CN116576629B

CN116576629B - System control method, system control device, electronic apparatus, and storage medium

Info

Publication number: CN116576629B
Application number: CN202310620058.5A
Authority: CN
Inventors: 童厚杰
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2023-05-29
Filing date: 2023-05-29
Publication date: 2024-03-12
Anticipated expiration: 2043-05-29
Also published as: CN116576629A

Abstract

The disclosure provides a system control method, a system control device, electronic equipment and a storage medium, relates to the technical field of artificial intelligence, and particularly relates to the technical fields of deep learning, machine learning, refrigeration technology and heating. The specific implementation scheme is as follows: according to at least one first candidate running state information of the target system, determining N levels of candidate equipment energy consumption influence information corresponding to the at least one candidate running state information, wherein the candidate equipment energy consumption influence information of the current level is associated with the candidate equipment energy consumption influence information of the previous level and the input information; determining target device control information from the current device control information of the current operating state information and the first candidate device control information of the at least one first candidate operating state information according to the first current selection information corresponding to the current operating state information and the first candidate selection information corresponding to the at least one first candidate operating state information; and controlling the operation of the target system according to the control information of the target equipment.

Description

System control method, system control device, electronic apparatus, and storage medium

Technical Field

The present disclosure relates to the field of artificial intelligence, and in particular, to the fields of deep learning, machine learning, refrigeration technology, and heating technology. And more particularly, to a system control method, a system control device, an electronic apparatus, and a storage medium.

Background

The target system may refer to a system that needs to generate a target amount while consuming energy. The target amount may refer to a physical amount for satisfying a predetermined demand. The predetermined demand may refer to a predetermined load. For example, the target system may include one of a refrigeration system and a heating system.

The energy consumption of the target system can be reduced by adjusting the device control parameters of the target system.

Disclosure of Invention

The present disclosure provides a system control method, a system control device, an electronic apparatus, and a storage medium.

According to an aspect of the present disclosure, there is provided a system control method including: determining N levels of candidate device energy consumption influence information corresponding to at least one candidate operation state information according to at least one first candidate operation state information of a target system, wherein the first candidate operation state information comprises first candidate device control information and current environment information, the candidate device energy consumption influence information of the current level is associated with candidate device energy consumption influence information of a previous level and input information, the input information of the previous level is used for determining the candidate device energy consumption influence information of the previous level, the input information of the previous level is determined according to the first candidate operation state information, and N is an integer greater than 1; determining target device control information meeting a predetermined energy saving condition from current device control information of the current operating state information and first candidate device control information of the at least one first candidate operating state information according to first current selection information corresponding to the current operating state information and first candidate selection information corresponding to the at least one first candidate operating state information, wherein the first candidate selection information comprises at least one hierarchy of the candidate device energy consumption influence information; and controlling the operation of the target system according to the target equipment control information.

According to another aspect of the present disclosure, there is provided a system control apparatus including: a first determining module, configured to determine N levels of candidate device energy consumption influence information corresponding to at least one first candidate operation state information of a target system according to the at least one first candidate operation state information, where the first candidate operation state information includes first candidate device control information and current environment information, the candidate device energy consumption influence information of the current level is associated with candidate device energy consumption influence information of a previous level and input information, the input information of the previous level is used to determine candidate device energy consumption influence information of the previous level, the input information of the previous level is determined according to the first candidate operation state information, and N is an integer greater than 1; a second determining module, configured to determine target device control information that meets a predetermined energy saving condition from current device control information of the current operation state information and first candidate device control information of the at least one first candidate operation state information according to first current selection information corresponding to the current operation state information and first candidate selection information corresponding to the at least one first candidate operation state information, where the first candidate selection information includes at least one hierarchy of the candidate device energy consumption influence information; and the control module is used for controlling the operation of the target system according to the control information of the target equipment.

According to another aspect of the present disclosure, there is provided an electronic device including: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform a method as described in the present disclosure.

According to another aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform a method as described in the present disclosure.

According to another aspect of the present disclosure, there is provided a computer program product comprising a computer program which, when executed by a processor, implements a method as described in the present disclosure.

It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification.

Drawings

The drawings are for a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:

FIG. 1 schematically illustrates an exemplary system architecture to which system control methods and apparatus may be applied, according to embodiments of the present disclosure;

FIG. 2 schematically illustrates a flow chart of a system control method according to an embodiment of the disclosure;

FIG. 3 schematically illustrates a schematic diagram of determining N levels of candidate device energy consumption impact information corresponding to candidate operating state information according to first candidate operating state information of a target system according to an embodiment of the disclosure;

FIG. 4A schematically illustrates an example schematic diagram of a refrigeration system according to an embodiment of the disclosure;

fig. 4B schematically illustrates a schematic diagram of determining N levels of candidate device energy consumption impact information corresponding to first candidate operating state information according to the first candidate operating state information in the case where the target system is a refrigeration system according to an embodiment of the present disclosure;

FIG. 5 schematically illustrates a block diagram of a system control device according to an embodiment of the disclosure; and

fig. 6 schematically illustrates a block diagram of an electronic device adapted to implement a system control method according to an embodiment of the disclosure.

Detailed Description

Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

Fig. 1 schematically illustrates an exemplary system architecture to which the system control methods and apparatuses may be applied according to embodiments of the present disclosure.

It should be noted that fig. 1 is only an example of a system architecture to which embodiments of the present disclosure may be applied to assist those skilled in the art in understanding the technical content of the present disclosure, but does not mean that embodiments of the present disclosure may not be used in other devices, systems, environments, or scenarios.

As shown in fig. 1, a system architecture 100 according to this embodiment may include a target system 101, an acquisition device 102, a gateway device 103, and an energy consumption management device 104.

The target system 101 may refer to a system that generates a physical quantity by consuming energy. The target system 101 may include at least one target device. For example, the target system 101 may include one of a refrigeration system, a heating system, a system having both a refrigeration function and a heating function, and the like. The physical quantity may include one of a cooling quantity and a heating quantity. A refrigeration system may refer to a system for generating refrigeration. A heating system may refer to a system for generating heat. The refrigeration system may comprise one of a data center (Internet Data Center, IDC) refrigeration system and a public building refrigeration system, etc. Public buildings may include at least one of buildings, airports, stations, subways, and the like. The predetermined load of the refrigeration system of the data center may include an IT (Internet Technology ) load.

For example, the target system 101 is a refrigeration system. The refrigeration system may include a cooling water system, a chilled water system, a refrigeration unit system, a transmission and distribution system, and the like. The cooling water system may be a system in which the heat-exchanged cooling water is recovered and recycled. The cooling water system may include at least one of a cooling pump, a cooling tower, a cooling water pipe, and the like. Chilled water systems may refer to systems in which the temperature of water is reduced by absorbing the cold energy of the refrigerant evaporating. The chilled water system may comprise at least one of the following target devices: a freeze pump, a water separator, a water collector, etc. The refrigeration unit system may include at least one of the following target devices: compressors, condensers, evaporators, and throttling elements, etc.

For example, the target system 101 is a heating system. The heating system may include a heat source, a thermal circulation system, and a heat dissipation device. The heat source may comprise a heat medium device. The thermal cycling system may include at least one of the following target devices: pipe network equipment, heat medium conveying equipment and the like. The low temperature heating medium is heated in a heat source and becomes a high temperature heating medium after absorbing heat. The high-temperature heating medium is conveyed to the terminal equipment through the heating medium conveying equipment, and the heat is released through the heat dissipation equipment, so that the temperature of the terminal equipment is increased. The high temperature heat medium is reduced in temperature after heat dissipation, and becomes a low temperature heat medium. The low-temperature heat medium returns to the heat source through the recovery pipeline for recycling. The temperature of the end device is maintained substantially constant by continuously transferring heat from the heat source to the end device to supplement the heat loss of the end device. The thermal cycling system may include at least one of the following target devices: primary pipe network, secondary pipe network, heat exchange station, etc. The heat source transmits the heat medium to the heat exchange station through the primary pipe network, and the heat exchange station transmits the heat medium to the terminal equipment through the secondary pipe network.

The acquisition device 102 may be used to acquire operational status information of the target system 101. For example, the collection device 102 may include at least one of a temperature sensor, a flow sensor, a pressure sensor, a smart meter, a smart water meter, and the like.

For example, the target system 101 is a refrigeration system. The temperature sensor may include at least one of a chilled side water inlet temperature sensor, a chilled side water outlet temperature sensor, a chilled water flow sensor, a cooling side water inlet temperature sensor, a cooling side water outlet temperature sensor, a cooling tower water inlet temperature, a cooling tower water outlet temperature, a dry bulb temperature sensor, a wet bulb temperature sensor, and the like. The flow sensor may include at least one of a refrigeration side flow sensor, a cooling water flow sensor, and the like.

For example, the target system 101 is a heating system. The temperature sensor may include at least one of a heat exchange station primary side water inlet temperature sensor, a heat exchange station primary side water outlet temperature sensor, a heat exchange station secondary side water inlet temperature sensor, a heat exchange station secondary side water outlet temperature sensor, and the like. The flow sensor may include at least one of a heat exchange station primary side flow sensor, a heat station secondary side flow sensor, and the like.

Gateway device 103 may be used as a transmission medium between acquisition device 102 and energy consumption management device 104. For example, the gateway device 103 includes at least one of an industrial personal computer, an edge gateway, and the like. Since in the production process, a plurality of sensors corresponding to the production lines can collect information in a time-sharing manner, if each sensor reports information to the energy consumption management device 104, in the case that the performance of the energy consumption management device 104 is low, since more information is received at the same time, a situation of refusing access occurs, and thus, the sensor information sent by the collection device 102 can be managed by the gateway device 103, and the information can be sent to the energy consumption management device 104.

For example, the sending of sensor information by the acquisition device 102 to the gateway device 103 may be implemented based on a high performance non-blocking communication framework. For example, a high performance non-blocking communication framework may include Netty. Gateway device 103 and acquisition device 102 may implement gateway device 103 to send control instructions to acquisition device 102 based on TCP (Transmission Control Protocol ). In addition, the gateway device 103 may also implement information transmission through the message queue and the energy consumption management device 104.

The energy consumption management device 104 may be various types of servers that provide various services. For example, the server 104 may be a cloud server. The cloud server is a host product in a cloud computing service system, and overcomes the defects of high management difficulty and weak service expansibility in the traditional physical host and VPS (Virtual Private Server ) service.

The acquisition device 102 may send the acquired operation state information to the gateway device 103, and the gateway device 103 may send the operation state information to the energy consumption management device 104. The energy consumption management apparatus 104 may be used to perform the system control method provided by the embodiments of the present disclosure. Accordingly, the system control device provided in the embodiments of the present disclosure may also be disposed in the energy consumption management apparatus 104.

In the case of reporting the running status information, the gateway device 103 may serve as a sender, package the running status information into a predetermined format, and push the running status information into a message queue. The energy consumption management device 104 may consume the operating state information from the message queue as a consumer and store the operating state information in the database. For example, the high performance requirement of the energy consumption management device 104 can be met by adopting a cache batch pushing mode, namely, the data pressure can be transferred to the gateway device 103 through batch pushing, and the instantaneity and usability of the energy consumption management device 104 can be ensured.

In the case of issuing control instructions, gateway device 103 may act as a consumer to retrieve control instructions from a message queue. The energy consumption management device 104 may act as a sender to send control instructions to the message queue. The mapping relationship between the gateway device 103 and the message queue can be configured by dividing the resource by the message queue, so that the mapping relationship can be positioned relatively quickly in case of abnormal problem.

It should be understood that the number of target systems, acquisition devices, gateway devices and energy consumption management devices in fig. 1 are merely illustrative. There may be any number of target systems, acquisition devices, gateway devices, and energy consumption management devices, as desired for implementation.

It should be noted that the sequence numbers of the respective operations in the following methods are merely representative of the operations for the purpose of description, and should not be construed as representing the order of execution of the respective operations. The method need not be performed in the exact order shown unless explicitly stated.

The inventive concept process of the present disclosure will be described with reference to the description of fig. 1, and some of the concepts related to the present disclosure will be described before explaining the inventive concept.

The operating state information may refer to information associated with the operating condition of the device. The information associated with the operating state of the device may refer to information affecting the operating condition of the device. The motion state information may include motion state dimension information of at least one dimension. For example, the operation state information may include at least one of device control information and environment information. The device control information may refer to information for controlling the operation of the target system. The environment information may refer to information of an environment in which the target system is located. The device control information may include device control dimension information of at least one dimension. The environment information may include environment dimension information of at least one dimension.

In the case where the target system is a refrigeration system, the device control dimension information for the at least one dimension may include at least one of: the temperature of the water inlet at the freezing side, the temperature of the water outlet at the freezing side, the temperature difference of the water outlet at the freezing side, the temperature of the water inlet at the cooling side, the temperature difference of the water outlet at the cooling side, the water inlet temperature of the cooling tower, the water outlet temperature of the cooling tower, the water inlet and outlet temperature difference of the cooling tower, the flow rate of the frozen water, the flow rate of the cooled water, the frequency of a frozen pump, the frequency of the cooling pump, the frequency of the fan of the cooling tower and the like. In the case where the target system is a heating system, the device control dimension information of at least one dimension may include at least one of: the water inlet temperature of the primary side of the heat exchange station, the water outlet temperature of the primary side of the heat exchange station, the primary side flow of the heat exchange station, the water inlet temperature of the secondary side of the heat exchange station, the water outlet temperature of the secondary side of the heat exchange station, the secondary side flow of the heat exchange station and the like. The environmental dimension information for the at least one dimension may include at least one of: temperature and humidity, etc. The temperature may include at least one of: dry bulb temperature and wet bulb temperature, etc.

In the process of realizing the inventive concept, it is found that, since the target system may include at least one target device, the device control information of the target system may include at least one dimensional device control dimension information, and the target device may have at least one device control dimension information corresponding to the target device, and there is a mutual coupling between the device control dimension information corresponding to each target device, it is necessary to optimize each device control dimension information by jointly considering the device control dimension information corresponding to each target device from the viewpoint of the device energy consumption information of the target system, so that the device energy consumption information of the target system can meet a predetermined energy saving condition in the case of controlling the operation of the target system according to the target device control information, and thus the energy saving effect of the target system is improved. The device energy consumption information may be determined based on device energy consumption sub-information of at least one predetermined device. The predetermined device may be one of the at least one target device.

In addition, it is further found that, since the target system is a circulation system from the primary side to the secondary side, information transmission between the respective target devices has a precedence relationship, and thus, there is a precedence relationship between device control dimension information corresponding to the respective target devices. For example, in the case where the energy consumption information of the refrigerating unit needs to be determined, the cooling water flow is determined according to the cooling pump frequency, then the cooling tower water inlet temperature and the cooling tower water outlet temperature are determined according to the cooling water flow, the cooling tower fan frequency, the dry bulb temperature and the wet bulb temperature, the cooling side water inlet temperature and the cooling side water outlet temperature are determined according to the cooling tower water inlet temperature and the cooling tower water outlet temperature, and then the energy consumption information of the refrigerating unit is finally determined according to the cooling side water inlet temperature, the cooling side water outlet temperature, the freezing side water inlet temperature and the freezing side water outlet temperature.

If each target device is considered as a whole, the above-mentioned precedence relationship is difficult to embody, and the energy-saving effect is reduced. For example, in the case where the cooling pump flow rate, the cooling side water inlet temperature, the cooling side water outlet temperature, and the refrigerating unit energy consumption information need to be determined, if the cooling pump frequency, the cooling tower fan frequency, the dry bulb temperature, the wet bulb temperature, the freezing side water inlet temperature, and the freezing side water outlet temperature are spliced, and the cooling side flow rate, the cooling side water inlet temperature, the cooling side water outlet temperature, and the refrigerating unit energy consumption information are determined based on the spliced information, the cooling side flow rate, the cooling side water inlet temperature, the cooling side water outlet temperature, and the refrigerating unit energy consumption information are synchronously output, and no precedence relation among the cooling side flow rate, the cooling side water inlet temperature, the cooling side water outlet temperature, and the refrigerating unit energy consumption information is represented.

Therefore, the embodiment of the disclosure introduces the concept of cascade into a process of jointly considering the device control dimension information corresponding to each target device, and proposes a system control scheme based on reasoning so as to improve the energy saving effect. For example, N levels of candidate device energy consumption impact information corresponding to at least one first candidate operating state information of the target system is determined based on the at least one first candidate operating state information. The first candidate operating state information includes first candidate device control information and current environment information. The candidate device energy consumption influence information of the current level is associated with the candidate device energy consumption influence information and the input information of the previous level. The input information of the previous hierarchy is used to determine candidate device energy consumption impact information of the previous hierarchy. The input information of the previous level is determined based on the first candidate operating state information.

Determining target device control information satisfying a predetermined power saving condition from the current device control information of the current operation state information and the first candidate device control information of the at least one first candidate operation state information according to the first current selection information corresponding to the current operation state information and the first candidate selection information corresponding to the at least one first candidate operation state information. The first candidate selection information includes candidate device energy consumption impact information of at least one hierarchy. And controlling the operation of the target system according to the control information of the target equipment.

According to the embodiment of the disclosure, since the candidate device energy consumption influence information of the current level is associated with the candidate device energy consumption influence information of the previous level and the input information, the input information of the previous level is used for determining the candidate device energy consumption influence information of the previous level, and the input information of the previous level is determined according to the first candidate running state information, the accuracy of the candidate device energy consumption influence information is improved. On the basis, according to the first current selection information corresponding to the current operation state information and the first candidate selection information corresponding to the at least one first candidate operation state information, the target device control information is determined from the current device control information of the current operation state information and the first candidate device control information of the at least one first candidate operation state information, and the first candidate selection information comprises at least one layer of candidate device energy consumption influence information, so that the accuracy of the target device control information is improved, and the target device control information meets the preset energy saving condition, so that the energy saving effect is improved.

Fig. 2 schematically illustrates a flow chart of a system control method according to an embodiment of the present disclosure.

As shown in fig. 2, the method 200 includes operations S210 to S230.

In operation S210, N levels of candidate device energy consumption influence information corresponding to at least one candidate operation state information are determined according to at least one first candidate operation state information of the target system.

In operation S220, target device control information satisfying a predetermined power saving condition is determined from the current device control information of the current operation state information and the first candidate device control information of the at least one first candidate operation state information according to the first current selection information corresponding to the current operation state information and the first candidate selection information corresponding to the at least one first candidate operation state information.

In operation S230, the target system is controlled to operate according to the target device control information.

According to an embodiment of the present disclosure, the first candidate operation state information may include first candidate device control information and current environment information. The candidate device energy consumption influence information of the current hierarchy may be associated with the candidate device energy consumption influence information of the previous hierarchy and the input information. The input information of the previous tier may be used to determine candidate device energy consumption impact information of the previous tier. The input information of the previous level may be determined based on the first candidate operating state information. N may be an integer greater than 1. N may be configured according to actual service requirements, which is not limited herein. For example, n=3.

According to an embodiment of the present disclosure, the target system may include one of a cooling system, a heating system, a system having both a cooling function and a heating function, and the like. The target system may comprise at least one target device. The target device may refer to a device that requires energy to be consumed. For example, where the target system is a refrigeration system, the at least one target device may include at least one of: cooling pump, cooling tower, plate heat exchanger, freeze pump, water knockout drum, water collector, compressor, condenser and evaporator etc.. In the case where the target system is a heating system, the at least one target device may comprise at least one of: heating media equipment, heat dissipation equipment and the like. In addition, the description of the target system and the target device may be referred to the corresponding parts above, and will not be repeated here.

According to an embodiment of the present disclosure, both the first candidate operation state information and the current operation state information may be candidate operation state information. The first candidate operating state information and the current operating state information may be used to determine target device control information. The target device may have first candidate operating state information. The first candidate operating state information may include first candidate device control information and current environment information. The current operating state information may include current device control information and current environment information. The first candidate device control information may refer to device control information corresponding to a non-current period. The current device control information may refer to device control information corresponding to a current period. The first candidate device control information may be determined from historical device control information corresponding to a non-current period. Alternatively, the first candidate device control information may be determined from a predetermined device control parameter range. Alternatively, the first candidate device control information may be determined from historical device control information corresponding to a non-current period and a predetermined device control parameter range. The current environmental information may refer to environmental information corresponding to the current period. The non-current period may refer to a period prior to the current period. The current time period may include at least one time. The non-current time period may include at least one time. The description of the first candidate device control information and the current device control information may be referred to above for the description of the device control information, which is not limited herein. The description of the current environmental information may be referred to above for the description of the environmental information, which is not limited herein.

According to embodiments of the present disclosure, candidate device energy consumption influence information may refer to information capable of influencing the energy consumption of the target system. Since at least two pieces of first candidate device control dimension information having an association relationship may exist among the plurality of pieces of first candidate device control dimension information included in the first candidate device control information, the at least two pieces of first candidate device control dimension information having an association relationship may refer to predetermined first candidate device control dimension information being determined from other first candidate device control dimension information. The other first candidate device control dimension information may refer to at least one first candidate device control dimension information other than the predetermined first candidate device control dimension information among the at least two first candidate device control dimension information having the association relationship, and thus, the first candidate device control information and the at least one hierarchy of candidate device energy consumption influence information may have information of the same dimension.

According to an embodiment of the present disclosure, the at least one hierarchy of candidate device energy consumption impact information may include candidate device energy consumption information. The candidate device energy consumption information may include candidate device energy consumption sub-information of at least one predetermined device. The hierarchy corresponding to the at least one candidate device energy consumption sub-information may be identical, partially identical, and completely different. The predetermined device may be one of the at least one target device.

According to an embodiment of the present disclosure, in case the target system is a refrigeration system, the at least one hierarchy of candidate device energy consumption impact information may further comprise at least one of: candidate cold energy, candidate cooling water flow, candidate cooling side water inlet temperature, candidate cooling side water outlet temperature, candidate freezing side flow, candidate freezing side water inlet temperature, candidate freezing side water outlet temperature, candidate cooling tower water inlet temperature, candidate cooling tower water outlet temperature and the like. Candidate coldness may refer to candidate demand information. In the case where the target system is a heating system, the at least one hierarchy of candidate device energy consumption impact information may further include at least one of: the heat exchange system comprises candidate heat, a candidate heat exchange station primary side water inlet temperature, a candidate heat exchange station primary side water outlet temperature, a candidate heat exchange station primary side flow, a candidate heat exchange station secondary side water inlet temperature, a candidate heat exchange station secondary side water outlet temperature, a candidate heat exchange station secondary side flow and the like. Candidate heat may refer to candidate demand information.

According to an embodiment of the present disclosure, the first current selection information may be used as one of the basis for determining the target device control information. The first candidate selection information may be used as one of the bases for determining the target device control information. The first current selection information may include at least one of: current equipment energy consumption information, current equipment operation mode information and the like. The first candidate selection information may include at least one hierarchy of candidate device energy consumption impact information. The at least one hierarchy of candidate device energy consumption impact information may further include at least one of: candidate equipment energy consumption information, candidate demand information and the like. In addition, the first candidate selection information may further include candidate device operation mode information. The candidate demand information may include one of a candidate cold amount and a candidate heat amount. The device operation mode information may refer to an operation mode of the device. For example, the device operating mode information may include at least one of: a double-working-condition main machine ice making mode, a double-working-condition main machine ice melting mode, a base-load main machine mode, a plate heat exchanger refrigerating mode and the like. The candidate device operation mode information may refer to device operation mode information of the target device corresponding to the first candidate device control information. The current device operation mode information may refer to device operation mode information of the target device corresponding to the current period. The target device may have candidate device operating mode information and current device operating mode information. For the description of the candidate device operation mode information and the current device operation mode information, reference may be made to the description of the device operation mode information above, which is not described herein.

According to an embodiment of the present disclosure, the information of at least one dimension included in the first current selection information may each have a corresponding priority. The information of at least one dimension included in the first candidate selection information may each have a corresponding priority. For example, the priority of the device operation mode information may be higher than the device energy consumption information, i.e. the priority of the current device operation mode information is higher than the priority of the current device energy consumption information. The priority of the candidate device operation mode information is higher than the priority of the candidate device energy consumption information. Alternatively, the priority of the device energy consumption information may be higher than the priority of the device operation mode information. I.e. the priority of the current device energy consumption information is higher than the priority of the current device operation mode information. The priority of the candidate device energy consumption information is higher than the priority of the candidate device operation mode information.

According to an embodiment of the present disclosure, the target device control information may refer to device control information determined from the current device control information and the at least one first candidate device control information to satisfy a predetermined power saving condition. The target device control information may be used as device control information corresponding to the next period. Under the condition that the target system is controlled to operate according to the target device control information, the target device energy consumption information meets the preset energy saving condition. The target device energy consumption information may refer to device energy consumption information corresponding to target device control information. For example, in the case where it is determined that the current device control information satisfies the predetermined power saving condition, the current device control information may be determined as the target device control information. The target device energy consumption information may refer to current device energy consumption information corresponding to current device control information. In the case where it is determined that there is first candidate device control information satisfying the predetermined power saving condition among the at least one first candidate device control information, the first candidate device control information satisfying the predetermined power saving condition may be determined as the target device control information. The target device energy consumption information may refer to candidate device energy consumption information corresponding to the first candidate device control information satisfying the predetermined power saving condition.

According to an embodiment of the present disclosure, it may be determined that the current device control information satisfies a predetermined power saving condition by one of: in the case that the current device energy consumption is less than or equal to the predetermined energy consumption threshold, it may be determined that the current device control information satisfies the predetermined energy saving condition. In case the current device energy consumption is smaller than any one of the at least one candidate device energy consumption, it may be determined that the current device control information satisfies the predetermined energy saving condition. In the case that the current device energy consumption is less than or equal to the predetermined energy consumption threshold and the current device energy consumption is less than any one of the at least one candidate device energy consumption, it may be determined that the current device control information satisfies the predetermined energy saving condition. The target amount corresponding to the current device control information may be greater than or equal to a predetermined amount. The current device energy consumption information may be characterized by the current device energy consumption. The candidate device energy consumption information may be characterized by candidate device energy consumption. The current demand information may be characterized by a target amount corresponding to the current device control information. The target amount corresponding to the current device control information may refer to a target amount generated by the target system in the case where the operation of the target system is controlled according to the current device control information. The predetermined amount may refer to a physical amount capable of satisfying a predetermined demand.

According to an embodiment of the present disclosure, it may be determined that the first candidate device control information satisfies the predetermined power saving condition by one of: in the case where the target amount corresponding to the first candidate device control information is greater than or equal to the predetermined amount, if the candidate device energy consumption is less than or equal to the predetermined energy consumption threshold, it may be determined that the first candidate device control information satisfies the predetermined energy saving condition. In the case where the target amount corresponding to the first candidate device control information is greater than or equal to the predetermined amount, if the candidate device power consumption is smaller than any one of the other device power consumption, it may be determined that the first candidate device control information satisfies the predetermined energy saving condition. In the case where the target amount corresponding to the first candidate device control information is greater than or equal to the predetermined amount, if the candidate device energy consumption is any one of less than or equal to the predetermined energy consumption threshold and less than the other device energy consumption, it may be determined that the first candidate device control information satisfies the predetermined energy saving condition. The candidate device energy consumption information may be characterized by candidate device energy consumption. The candidate demand information may be characterized by a target amount corresponding to the first candidate device control information. The target amount corresponding to the first candidate device control information may refer to a target amount generated by the target system in the case where the operation of the target system is controlled according to the first candidate device control information. Other device energy consumption information may be characterized by other device energy consumption. The other device energy consumption information may include any one of the current device energy consumption information and the at least one candidate device energy consumption information other than the candidate device energy consumption information.

According to embodiments of the present disclosure, in the case where N > 1, in response to 1 < n+.ltoreq.N, the N-th level candidate device energy consumption impact information may be associated with the N-1-th level candidate device energy consumption impact information and the N-1-th level input information. For example, the n-th level candidate device energy consumption influence information may be determined from the n-th level input information. The input information according to the n-th hierarchy may be determined according to the n-1-th hierarchy's candidate device energy consumption influence information and the n-1-th hierarchy's input information. The n-1 th level candidate device energy consumption influence information and the n-1 th level input information may be determined according to the first candidate operation state information. The nth level may characterize the current level. The n-1 th level may characterize the previous level. In the case of n=1 or in the case of N > 1, the candidate device energy consumption influence information of the 1 st hierarchy may be determined from the input information of the 1 st hierarchy in response to n=1. The level 1 input information may be determined based on the first candidate operating state information. For example, the first candidate operating state information may be determined as the level 1 input information.

According to the embodiment of the disclosure, N levels of candidate device energy consumption influence information corresponding to each of at least one first candidate operation state information of a target system may be determined according to the at least one first candidate operation state information. For example, for a first candidate operation state information of the at least one first candidate operation state information, in the case where N > 1, in response to 1 < n.ltoreq.N, the candidate device energy consumption influence information of the nth level is determined from the candidate device energy consumption influence information of the nth-1 level and the input information of the nth-1 level. In the case of n=1 or in the case of N > 1, in response to n=1, the input information of the 1 st hierarchy is determined from the first candidate operation state information. And determining the candidate equipment energy consumption influence information of the 1 st level according to the input information of the 1 st level. When the candidate device energy consumption influence information corresponding to each hierarchy level is determined, N hierarchy level candidate device energy consumption influence information corresponding to the first candidate operation state information may be obtained according to the 1 st hierarchy level candidate device energy consumption influence information, the 2 nd hierarchy level candidate device energy consumption influence information, the … … nth hierarchy level candidate device energy consumption influence information, the … … nth-1 st hierarchy level candidate device energy consumption influence information, and the nth hierarchy level candidate device energy consumption influence information.

According to an embodiment of the present disclosure, target device control information satisfying a predetermined power saving condition may be determined from current device control information corresponding to current operation state information and first candidate device control information corresponding to at least one first candidate operation state information, respectively, according to the first current selection information and the at least one first candidate selection information. For example, in the case where the first current selection information includes current device energy consumption information and the first candidate selection information corresponding to each of the at least one first candidate operation state information includes candidate device energy consumption information, target device control information satisfying a predetermined energy saving condition is determined from the current device control information and the at least one first candidate device control information based on the current device energy consumption information corresponding to the current device control information and the candidate device energy consumption information corresponding to each of the at least one first candidate device control information.

Alternatively, in the case where the first current selection information includes current device energy consumption information, the first candidate selection information corresponding to each of the at least one first candidate operation state information includes candidate device energy consumption information and candidate demand information, at least one second candidate operation state information is determined from the at least one first candidate operation state information according to the candidate demand information corresponding to each of the at least one first candidate operation state information. And determining target equipment control information from the current equipment control information and the first candidate equipment control information of the at least one second candidate running state information according to the current equipment energy consumption information corresponding to the current equipment control information and the candidate equipment energy consumption information corresponding to the first candidate equipment control information of the at least one second candidate running state information.

Alternatively, in the case that the first current selection information includes current device energy consumption information and current device operation mode information, the first candidate selection information corresponding to each of the at least one first candidate operation state information includes candidate device energy consumption information and candidate device operation mode information, at least one fourth candidate operation state information is determined from the at least one first candidate operation state information according to the current device operation mode information corresponding to the current operation state information and the candidate device operation mode information corresponding to the at least one first candidate operation state information. And determining target equipment control information from the current equipment control information and the first candidate equipment control information of the at least one fourth candidate running state information according to the current equipment energy consumption information corresponding to the current equipment control information and the candidate equipment energy consumption information corresponding to the first candidate equipment control information of the at least one fourth candidate running state information.

Alternatively, in the case where the first current selection information includes current device energy consumption information and current device operation mode information, the first candidate selection information corresponding to each of the at least one first candidate operation state information includes candidate device energy consumption information, candidate demand information, and candidate device operation mode information, at least one second candidate operation state information is determined from among the at least one first candidate operation state information according to the candidate demand information corresponding to each of the at least one first candidate operation state information. At least one third candidate operation state information is determined from the at least one second candidate operation state information according to the current device operation mode information corresponding to the current operation state information and the candidate device operation mode information corresponding to the at least one second candidate operation state information. And determining target equipment control information from the current equipment control information and the first candidate equipment control information of the at least one third candidate running state information according to the current equipment energy consumption information corresponding to the current equipment control information and the candidate equipment energy consumption information corresponding to the first candidate equipment control information of the at least one third candidate running state information.

According to the embodiment of the present disclosure, after the target device control information is determined, the target system may be controlled to operate according to the target device control information so that the energy consumption of the target system can satisfy a predetermined energy saving condition.

According to an embodiment of the present disclosure, the first candidate run state information may include first candidate run state dimension information of at least one dimension.

According to an embodiment of the present disclosure, the above-described system control method may further include the following operations.

And obtaining the historical running state characteristic information corresponding to the plurality of candidate dimensions according to the historical running state information corresponding to the plurality of candidate dimensions. And determining importance degrees corresponding to the plurality of candidate dimensions according to the historical operation state characteristic information corresponding to the plurality of candidate dimensions. At least one dimension is determined from the plurality of candidate dimensions according to importance levels corresponding to the plurality of candidate dimensions.

According to an embodiment of the present disclosure, the historical operating state information may refer to operating state information corresponding to a first historical period. The first history period may refer to a period preceding the current period. The importance of the candidate dimension may characterize the importance of the candidate dimension. The relationship between importance and importance level may be configured according to actual service requirements, and is not limited herein. For example, the smaller the importance of the candidate dimension, the greater the importance of the candidate dimension. Alternatively, the greater the importance of the candidate dimension. Alternatively, the smaller the importance of the candidate dimension. Alternatively, the greater the importance of the candidate dimension, the lesser the importance of the candidate dimension.

According to the embodiment of the disclosure, the plurality of candidate dimensions can be clustered according to the historical operation state characteristic information corresponding to each of the plurality of candidate dimensions, so as to obtain clustering information. The cluster information may include at least one cluster and categories corresponding to each of the at least one cluster. For a cluster of the at least one cluster, the most number of the at least one class corresponding to the cluster may be determined as the class corresponding to the cluster. Clusters can be characterized by a cluster center. And determining importance degrees corresponding to the candidate dimensions according to the clustering information. For example, a degree of distinction corresponding to each of the plurality of candidate dimensions may be determined from the cluster information. The degree of distinction corresponding to each of the plurality of candidate dimensions is determined as the degree of importance corresponding to each of the plurality of candidate dimensions. The degree of differentiation of the candidate dimension may characterize the degree of differentiation of the candidate dimension in the categories corresponding to each of the at least one cluster. For a candidate dimension in the at least one candidate dimension, historical operation state characteristic information corresponding to the candidate dimension and distance information between the historical operation state characteristic information and at least one clustering center can be determined, and at least one distance information corresponding to the candidate dimension is obtained. And determining the distinguishing degree corresponding to the candidate dimension according to at least one piece of distance information corresponding to the candidate dimension.

According to the embodiment of the disclosure, historical running state characteristic information corresponding to each of a plurality of candidate dimensions is obtained according to the historical running state information corresponding to each of the plurality of candidate dimensions. For example, for a candidate dimension of the plurality of candidate dimensions, feature extraction may be performed on the historical operating state information corresponding to the candidate dimension, resulting in historical operating state feature information corresponding to the candidate dimension.

According to the embodiment of the disclosure, the plurality of candidate dimensions may be ranked according to the importance degrees corresponding to the plurality of candidate dimensions, so as to obtain the first ranking information. At least one dimension is determined from the plurality of candidate dimensions based on the first ordering information. Alternatively, at least one dimension is determined from the plurality of candidate dimensions according to a predetermined importance and an importance corresponding to each of the plurality of candidate dimensions. And comparing the preset importance with the importance corresponding to the candidate dimension in at least one candidate dimension to obtain comparison information. Based on the comparison information, it is determined whether the candidate dimension is a dimension. The predetermined importance level may be configured according to actual service requirements, and is not limited herein.

For example, the greater the importance of the candidate dimension. The plurality of candidate dimensions may be ranked in order of importance of the candidate dimensions from greater than lesser, and a first predetermined number of candidate dimensions from the plurality of candidate dimensions that are top ranked may be determined as the at least one dimension. Alternatively, for a candidate dimension among a plurality of candidate dimensions, in a case where it is determined that the importance degree corresponding to the candidate dimension is greater than or equal to a predetermined importance degree, the candidate dimension may be determined as a dimension. The first predetermined number may be configured according to actual service requirements, and is not limited herein.

According to the embodiment of the disclosure, since at least one dimension is determined from the plurality of candidate dimensions according to the importance degrees corresponding to the plurality of candidate dimensions, the data size of the first candidate operation state information is reduced, resource overhead is saved, and furthermore, since the first candidate operation state dimension information of each dimension included in the first candidate operation state information is important and noise-removed information, the data quality of the first candidate operation state information is improved, and the accuracy of the target device control information is further improved.

According to an embodiment of the present disclosure, obtaining historical operating state characteristic information corresponding to a plurality of candidate dimensions from the historical operating state information corresponding to the plurality of candidate dimensions may include the following operations.

And processing the historical operation state information corresponding to the plurality of candidate dimensions by using the characterization model to obtain the historical operation state characteristic information corresponding to the plurality of candidate dimensions.

According to an embodiment of the present disclosure, in contrast learning, a child sample obtained by data enhancement of a parent sample is considered as a positive sample for the parent sample, because the child sample and the parent sample are the same in category, maintaining the same semantic information as each other. A parent sample may refer to a sample that is the subject of data enhancement processing. For the same parent sample, multiple data enhancements may be performed on the parent sample, resulting in multiple child samples. Negative samples may refer to other samples that are of a different class than the parent sample.

According to embodiments of the present disclosure, the characterization model may be derived by training a self-supervision model using the loss function values. The loss function value may be determined based on the loss function from sample operation state characteristic information of the positive sample and sample operation state characteristic information of a plurality of negative samples corresponding to the positive sample. For example, the characterization model may be a self-supervised model trained in the event that a predetermined end condition is met. The sample operation state feature information of the positive sample and the sample operation state feature information of each of the plurality of negative samples corresponding to the positive sample may be input into a loss function to obtain a loss function value. Model parameters of the self-supervision model may be adjusted according to the loss function value until a predetermined end condition is met. The predetermined end condition may include at least one of: the loss function value converges and reaches the maximum training round. The loss function may include at least one of: infoNCE (Info Noise-contrastive Estimation, information Noise contrast estimation) and NCE (Noise-Constrastive Estimation Loss, noise contrast estimation), and the like. The loss function may further include a loss function obtained by improving the loss function.

According to an embodiment of the present disclosure, the self-supervising model may include at least one of: CPC (Contrastive Predictive Coding), AMDIM (Augmented Multiscale Deep InfoMax), MOCO (MOmentum COntrast ), simCLR (Simple Framework fOr Contrastive Learning of Visual Representations) and BYOL (Bootstrap Your Own Latent), etc.

For example, the self-supervising model may be a MOCO-based model. A momentum queue may refer to a queue having a certain length. A queue may include a plurality of queue elements. The plurality of queue elements are sequential, i.e., are queued in sequential order. The queue has the characteristic of first-in first-out, namely if new queue elements need to be added to the queue, the earliest enqueued queue elements can be dequeued and the new queue elements are added to the queue under the condition that the queue is full. Queue elements in the momentum queue may be referred to as sample run state characteristic information. The momentum queue may include a plurality of sample run state characteristic information. The sample operation state characteristic information included in the momentum queue may refer to sample operation state characteristic information corresponding to a negative sample. The momentum queue includes sample run state characteristic information that can be updated dynamically, i.e., each round has a momentum queue corresponding to that round. Updating the momentum queue corresponding to the current round is to add the sample running state characteristic information of the parent sample corresponding to the previous round to the momentum queue corresponding to the previous round, and remove one sample running state characteristic information of the momentum queue corresponding to the previous round from the queue according to a time sequence order, so that the number of the sample running state characteristic information included in the momentum queue is kept unchanged.

According to an embodiment of the present disclosure, a self-supervising model may include a first encoder and a second encoder. Multiple rounds of training may be performed on the first encoder and the second encoder until a predetermined end condition is met. The trained second encoder is determined as a characterization model.

According to an embodiment of the present disclosure, training the first encoder and the second encoder multiple times may include: and processing the parent sample corresponding to the current round by using a first encoder corresponding to the current round to obtain sample running state characteristic information of the parent sample corresponding to the current round. And processing the positive samples corresponding to the current round by using a second encoder corresponding to the current round to obtain sample running state characteristic information of the positive samples corresponding to the current round. Positive samples are obtained by data enhancement of parent samples. Based on the loss function, training a first encoder and a second encoder corresponding to the current run using sample run state feature information of a parent sample, sample run state feature information of a positive sample, and sample run state feature information of a plurality of negative samples corresponding to the current run. Sample operation state characteristic information of a plurality of negative samples corresponding to the current round is obtained according to the momentum queue corresponding to the current round and sample operation state characteristic information of a parent sample based on a sample selection strategy corresponding to the current round.

According to the embodiment of the disclosure, the characterization model is obtained by training the self-supervision model by utilizing the sample running state characteristic information of the positive sample and the sample running state characteristic information of each of the plurality of negative samples corresponding to the positive sample, and the categories of the positive sample and the negative sample are different, so that the characterization model can learn more accurate running state characteristic information, the characterization effect of the characterization model is improved, and the accuracy of obtaining the sample running state characteristic information by utilizing the characterization model is improved.

According to an embodiment of the present disclosure, the plurality of negative samples corresponding to the positive sample may be determined from the plurality of candidate negative samples according to sample operation state characteristic information of the positive sample and sample operation state characteristic information of the plurality of candidate negative samples corresponding to the positive sample. The sample operation state characteristic information of the positive sample can be obtained by processing the positive sample by using a self-supervision model. The sample operating state characteristic information of the negative sample may be obtained by processing the negative sample using a self-supervision model.

According to the embodiments of the present disclosure, a plurality of negative samples may be determined from a plurality of candidate negative samples according to a similarity between sample operation state characteristic information of a positive sample and sample operation state characteristic information of the plurality of candidate negative samples corresponding to the positive sample. For example, the similarity between the sample operation state feature information of the positive sample and the sample operation state feature information of each of the plurality of candidate negative samples may be determined, so as to obtain a plurality of first similarities. A plurality of negative samples is determined from the plurality of candidate negative samples based on the first predetermined similarity threshold and the plurality of first similarities. The first similarity may characterize a degree of similarity between the positive sample and the candidate negative sample. The relationship between the first similarity and the degree of similarity may be configured according to actual service requirements, which is not limited herein. For example, the greater the first degree of similarity, the higher the degree of similarity. For a candidate negative sample of the plurality of candidate negative samples, the candidate negative sample may be determined as a negative sample in a case where it is determined that a first similarity between the sample operation state feature information of the positive sample and the sample operation state feature information of the candidate negative sample is less than or equal to a first predetermined similarity threshold. The first predetermined similarity threshold may be configured according to actual service requirements, and is not limited herein. The similarity may be configured according to actual service requirements, which is not limited herein. For example, the similarity may include at least one of: a literal similarity-based method, a text similarity-based method, an entity similarity-based method, and the like. The method based on literal similarity may include at least one of: euclidean distance, edit distance, dice coefficient, and Jaccard similarity, etc. The text similarity based method may include at least one of: cosine similarity, relative entropy, KL (Kullback-Leibler, KL) divergence, probability model similarity, and the like. The entity similarity based method may comprise at least one of: cluster-based methods and aggregation-based methods.

According to the embodiment of the disclosure, since the plurality of negative samples corresponding to the positive sample can be determined from the plurality of candidate negative samples according to the sample running state characteristic information of the positive sample and the sample running state characteristic information of the plurality of candidate negative samples corresponding to the positive sample, the difference between the negative sample corresponding to the positive sample and the positive sample is large, and therefore, the negative sample with small difference from the positive sample is effectively avoided from participating in the training of the model, and therefore, the probability of overfitting of the self-supervision model in the training stage is reduced.

And processing the plurality of second historical equipment control information of the target system based on the preprocessing method to obtain a plurality of first historical equipment control information of the target system.

According to an embodiment of the present disclosure, the pretreatment method may include at least one of: a data filtering method, a missing value processing method, a discretization processing method, a data standardization processing method and the like. The data filtering method may include an outlier processing method. Due to acquisition errors and the like, noise information and missing values are caused to exist in the second historical equipment control information. Since the presence of noise information and missing values will affect subsequent data mining analysis, the second historical device control information is required for data filtering and missing value processing. In addition, since there are second history device control dimension information of different magnitudes, data normalization processing can be performed on the second history device control information.

According to an embodiment of the present disclosure, the missing value processing method may include at least one of: a data deleting method and a data filling method. The data population method may comprise at least one of: statistical value filling method, K-neighbor filling method and interpolation method. In the case where the second history device control information is continuous data, the statistical value may be at least one of an average value of the second history device control information and a median of the second history device control information. In the case where the second historical device control information is discrete data, the statistical value may be a mode of the second historical device control information. The K-nearest neighbor filling method may be to determine nearest neighbor features having a second similarity to the missing-valued feature (i.e., missing feature) greater than or equal to a second predetermined similarity threshold, and then fill in the missing feature by weighting the nearest neighbor features according to the second similarity between the nearest neighbor features and the missing feature. The interpolation method may include a linear interpolation method. The discretization method may include a one-hot encoding method. The data standardization processing method can refer to a method for converting the value range of the dimension information of each level into the same level. The data normalization processing method can enable dimension information of various orders to be comparable in value. The data normalization processing method may include a zero-mean normalization processing method.

According to an embodiment of the present disclosure, second historical device control information satisfying a predetermined outlier condition is determined from a plurality of second historical device control information of a target system, resulting in at least one third historical device control information. And deleting the third historical device control information by a data deleting method in response to the absence value being a continuous plurality of values of a predetermined period of time in a case where it is determined that the absence value exists for the third historical device control information with respect to the third historical device control information of the at least one third historical device control information. And processing the third historical equipment control information by using a data filling method to obtain fourth historical equipment control information in response to the missing value being at least one discrete value. For example, in response to the missing value being at least one discrete value, the third historical device control information is processed using a linear interpolation method to obtain fourth historical device control information. And processing the fourth historical equipment control information by using a discretization processing method to obtain the first historical equipment control information. Alternatively, the fourth historical equipment control information is processed by using a data standardization processing method to obtain the first historical equipment control information.

According to the embodiment of the disclosure, the first historical equipment control information is obtained by preprocessing the second historical equipment control information, so that the accuracy of the first historical equipment control information is improved.

At least one second candidate device control information is determined from a plurality of historical device control information for the target system. First candidate device control information of the at least one first candidate operating state information is determined according to the at least one second candidate device control information.

According to an embodiment of the present disclosure, the history device control information may refer to device control information corresponding to the second history period. The second history period may refer to a period preceding the current period. The second historical period may be configured according to actual business requirements, and is not limited herein.

According to embodiments of the present disclosure, an occurrence time of a predetermined instruction may be determined in response to detecting the predetermined instruction. And determining a second history period according to the occurrence time of the predetermined instruction. A plurality of historical device control information corresponding to the second historical period is determined from the set of device control information. The predetermined instruction may refer to an instruction for triggering a system control task. The system control task may refer to a task for determining target device control information corresponding to a next period of time. The second history period may refer to a period prior to the occurrence time. The occurrence time may or may not belong to the current period.

For example, a plurality of times before the occurrence time may be determined, resulting in a second history period. A plurality of initial historical device control information corresponding to the second historical period is determined from the set of device control information. The plurality of initial historical device control information may be aggregated at a predetermined granularity to obtain a plurality of historical device control information. The plurality of historical device control information each has a corresponding time. The predetermined granularity may be configured according to actual service requirements, and is not limited herein. For example, the predetermined particle size may be one of minutes, hours, days, and the like.

According to an embodiment of the present disclosure, at least one second candidate device control information may be determined from a plurality of historical device control information based on predetermined selection information. The value range of the second candidate device control dimension information corresponding to each of the at least one dimension may be determined according to the at least one second candidate device control information. And determining at least one first candidate device control information according to the value range of the second candidate device control dimension information corresponding to the at least one dimension. For example, for a dimension of the at least one dimension, the at least one first candidate device control dimension information corresponding to the dimension may be determined according to a range of values of the second candidate device control dimension information corresponding to the dimension. At least one first candidate device control information is determined based on at least one first candidate device control dimension information corresponding to each of the at least one dimension.

According to the embodiment of the present disclosure, since the first candidate device control information of the at least one first candidate operation state information is determined according to the at least one second candidate device control information, which is determined from a plurality of history device control information of the target system, the history device control information is reliable information, not only the number of the first candidate device control information is reduced, but also the reliability of the first candidate device control information is improved. On the basis, the resource consumption is reduced, and the reliability of the control information of the target equipment is improved. In addition, due to the reduction of the number of the first candidate device control information, the determination efficiency of the target device control information is improved, and therefore the timely pushing requirement of the target device control information can be met.

According to an embodiment of the present disclosure, determining at least one second candidate device control information from among a plurality of historical device control information of a target system may include the following operations.

At least one second candidate device control information is determined from the plurality of historical device control information based on the second current selection information corresponding to the current device control information and the second candidate selection information corresponding to the plurality of historical device control information of the target system.

According to an embodiment of the present disclosure, the second current selection information may be used as one of the basis for determining the second candidate device control information. The second candidate selection information may be used as one of the bases for determining the second candidate device control information. The second current selection information may include at least one of: current load information, current environment information, and the like. The current load information may be derived from historical load information corresponding to the third historical period. For example, the current load information may be obtained by processing the historical load information corresponding to the third historical period using the load prediction model. The second candidate selection information may include historical load information and historical environment information. The current load information may refer to load information corresponding to the current period. The historical load information may refer to load information corresponding to the second historical period. The load information may include IT loads.

According to an embodiment of the present disclosure, in a case where the second current selection information includes current load information and the second candidate selection information includes historical load information, at least one third candidate device control information is determined from the plurality of historical device control information according to the current load information and the historical load information corresponding to each of the plurality of historical device control information. At least one third candidate device control information is determined as at least one first candidate device control information.

Alternatively, in the case where the second current selection information includes current environment information and the second candidate selection information includes historical environment information, at least one fourth candidate device control information is determined from the plurality of historical device control information based on the current environment information and the historical environment information corresponding to each of the plurality of historical device control information. At least one fourth candidate device control information is determined as at least one first candidate device control information.

Alternatively, in the case where the second current selection information includes current load information and current environment information, and the second candidate selection information includes historical load information and historical environment information, at least one fourth candidate device control information is determined from the plurality of historical device control information according to the current environment information and the historical environment information corresponding to each of the plurality of historical device control information. At least one first candidate device control information is determined from the at least one fourth candidate device control information based on the current load information and historical load information corresponding to each of the at least one fourth candidate device control information.

Alternatively, in the case where the second current selection information includes current load information and current environment information, and the second candidate selection information includes historical load information and historical environment information, at least one third candidate device control information is determined from the plurality of historical device control information according to the current load information and the historical load information corresponding to each of the plurality of historical device control information. At least one first candidate device control information is determined from the at least one third candidate device control information based on the current environment information and the historical environment information corresponding to each of the at least one third candidate device control information.

According to an embodiment of the present disclosure, the second current selection information may include current load information and current environment information. The second candidate selection information may include historical load information and historical environment information.

According to an embodiment of the present disclosure, determining at least one second candidate device control information from among a plurality of historical device control information according to second current selection information corresponding to current device control information and second candidate selection information corresponding to the plurality of historical device control information of the target system may include the following operations.

At least one third candidate device control information is determined from the plurality of historical device control information based on the current load information corresponding to the current device control information and the historical load information corresponding to the plurality of historical device control information of the target system. At least one second candidate device control information is determined from the at least one third candidate device control information based on the current environment information corresponding to the current device control information and the historical environment information corresponding to the at least one third candidate device control information.

According to an embodiment of the present disclosure, first difference information between current load information and historical load information corresponding to historical device control information is determined for the historical device control information of a plurality of historical device control information. And determining the historical device control information as third candidate device control information in the case that the first difference information meets the first preset difference condition. The first predetermined differential condition may be determined based on current load information. For example, the current load information may be characterized by the current load. The historical load information may be characterized by a historical load. The first difference information may be characterized by a first difference value. And determining the historical device control information as third candidate device control information in the case that the first difference value is less than or equal to the first predetermined difference threshold. For example, the first predetermined variance threshold may be determined based on a predetermined coefficient and a quarter bit distance of the current load value. For example, the first predetermined variance threshold may be a product of a predetermined system and a quarter bit distance of the current load value. The predetermined coefficient may be configured according to actual service requirements, and is not limited herein. For example, the predetermined coefficient may be 1.5.

According to an embodiment of the present disclosure, at least one second candidate device control information may be determined from at least one third candidate device control information according to the current environment information and the plurality of historical environment information. For example, for a third candidate device control information of the at least one third candidate device control information, second difference information between the current environment information and historical environment information corresponding to the third candidate device control information may be determined. In the case where it is determined that the second difference information satisfies the second predetermined difference condition, the third candidate device control information is determined as the first candidate device control information. For example, the second difference information may be characterized by a second difference value. And determining the third candidate device control information as the first candidate device control information in the case that the second difference value is determined to be less than or equal to the second predetermined difference threshold. For example, the second predetermined difference threshold may be configured according to actual service requirements, which is not limited herein

According to an embodiment of the present disclosure, determining at least one second candidate device control information from among the at least one third candidate device control information according to current environment information corresponding to the current device control information and historical environment information corresponding to the at least one third candidate device control information may include the following operations.

And determining the similarity between the current environment information corresponding to the current device control information and the historical environment information corresponding to the at least one third candidate device control information, and obtaining at least one similarity. At least one second candidate device control information is determined from the at least one third candidate device control information according to the at least one similarity.

According to an embodiment of the present disclosure, a third similarity between the current environmental information and at least one of the historical environmental information is determined, and at least one third similarity is obtained. The third similarity may characterize a degree of similarity between the current environmental information and the historical environmental information. The relationship between the third similarity and the similarity degree may be configured according to the actual service requirement, which is not limited herein. And ordering the at least one third candidate device control information according to the at least one third similarity to obtain second ordering information. At least one second candidate device control information is determined from the at least one third candidate device control information according to the second ordering information. Alternatively, at least one second candidate device control information may be determined from the at least one third candidate device control information according to a third predetermined similarity threshold and the at least one third similarity. The third predetermined similarity threshold may be configured according to actual service requirements, and is not limited herein. For example, the third similarity may be a number greater than 0 and less than or equal to 1. The third similarity may be a euclidean distance based similarity.

For example, the greater the third degree of similarity, the greater the degree of similarity. The third candidate device control information corresponding to each of the at least one third similarity may be sorted in order of the third similarity from the higher order to the lower order, and a second predetermined number of third candidate device control information, which is sorted earlier, may be determined from the at least one third candidate device control information as the at least one second candidate device control information. Alternatively, for the third candidate device control information of the at least one third candidate device control information, in a case where it is determined that the third similarity corresponding to the third candidate device control information is greater than or equal to the third predetermined similarity threshold, the third candidate device control information may be determined as the second candidate device control information. The second predetermined number may be configured according to actual service requirements, and is not limited herein.

According to the embodiment of the disclosure, since the history device control information is reliable information, determining at least one second device control information from a plurality of history device control information based on the current environment information and the history environment information and based on the current load information and the history load information, not only reduces the number of second candidate device control information, but also improves the reliability of the second candidate device control information. On the basis, the resource consumption is reduced, and the reliability of the control information of the target equipment is improved. In addition, due to the reduction of the number of the second candidate device control information, the determination efficiency of the target device control information is improved, and therefore the timely pushing requirement of the target device control information can be met.

According to an embodiment of the present disclosure, operation S210 may include repeatedly performing the following operations until the nth candidate device energy consumption influence information is obtained:

and determining nth fusion information corresponding to the at least one first candidate operation state information according to the candidate operation state characteristic information, the nth-1 fusion information and the nth-1 candidate equipment energy consumption influence information corresponding to the at least one first candidate operation state information of the target system. And obtaining the energy consumption influence information of the nth candidate equipment corresponding to the at least one first candidate running state information according to the nth fusion information corresponding to the at least one first candidate running state information.

According to an embodiment of the present disclosure, the candidate operation state characteristic information corresponding to the first candidate operation state information may be determined according to the first candidate operation state information. N may be an integer greater than 1 and less than or equal to N. The nth candidate device energy consumption impact information may characterize the nth hierarchical candidate device energy consumption impact information. The n-1 th fusion information may characterize the n-1 th hierarchical fusion information. The nth fusion information may characterize the nth hierarchical fusion information. n.epsilon. {2, … …, N-1, N }.

According to the embodiment of the disclosure, the at least one first candidate operation state information may be processed to obtain candidate operation state feature information corresponding to the at least one first candidate operation state information. For example, the at least one first candidate operation state information may be processed to obtain candidate operation state feature information corresponding to each of the at least one first candidate operation state information. For example, the time sequence feature extraction may be performed on the at least one first candidate operation state information, so as to obtain candidate operation state feature information corresponding to the at least one first candidate operation state information. The time sequence feature extraction may be performed on the at least one first candidate operation state information, so as to obtain candidate operation state feature information corresponding to each of the at least one first candidate operation state information.

According to the embodiment of the disclosure, in response to 1 < N being less than or equal to N, for first candidate operation state information in at least one first candidate operation state information, fusion processing can be performed on candidate operation state feature information, N-1 th fusion information and N-1 th candidate equipment energy consumption influence information corresponding to the first candidate operation state information to obtain N-th fusion information corresponding to the first candidate operation state information. And processing the nth fusion information corresponding to the first candidate operation state information to obtain nth candidate equipment energy consumption influence information corresponding to the first candidate operation state information. In response to n=1, the 1 st candidate device energy consumption influence information corresponding to the first candidate operation state information may be obtained from the candidate operation state characteristic information corresponding to the first candidate operation state information.

According to the embodiment of the disclosure, in the case of n=1, for the first candidate operation state information in the at least one first candidate operation state information, according to the candidate operation state feature information corresponding to the first candidate operation state information, the 1 st candidate device energy consumption influence information corresponding to the first candidate operation state information is obtained.

The system control method shown in fig. 2 is further described below with reference to fig. 3, 4A and 4B in conjunction with specific embodiments.

Fig. 3 schematically illustrates a schematic diagram of determining N levels of candidate device energy consumption impact information corresponding to candidate operating state information according to first candidate operating state information of a target system according to an embodiment of the present disclosure.

As shown in fig. 3, candidate operation state characteristic information is obtained according to the first candidate operation state information. And obtaining the energy consumption influence information of the 1 st candidate equipment according to the candidate running state characteristic information. And obtaining the 1 st fusion information according to the first candidate running state information, the candidate running state characteristic information and the 1 st candidate equipment energy consumption influence information.

And obtaining the 2 nd fusion information according to the 1 st fusion information, the 1 st candidate equipment energy consumption influence information and the candidate running state characteristic information. And obtaining the energy consumption influence information of the 2 nd candidate equipment according to the 2 nd fusion information. … …. And obtaining the nth fusion information according to the nth-1 fusion information, the nth-1 candidate equipment energy consumption influence information and the candidate running state characteristic information. And according to the nth fusion information, obtaining the energy consumption influence information of the nth candidate equipment. … …. And obtaining the N fusion information according to the N-1 fusion information, the N-1 candidate equipment energy consumption influence information and the candidate running state characteristic information. And according to the N fusion information, obtaining the energy consumption influence information of the N candidate equipment.

And extracting time sequence characteristics of at least one first candidate operation state information to obtain candidate operation state characteristic information corresponding to the at least one first candidate operation state information.

According to an embodiment of the present disclosure, the candidate operation state feature information corresponding to the first candidate operation state information may be obtained by performing time sequence feature extraction on the first candidate operation state information. For example, the at least one first candidate operating state information may be processed using the timing model to obtain candidate operating state characteristic information corresponding to the at least one first candidate operating state information. For example, the at least one first candidate operating state information is processed by using the time sequence model, so that candidate operating state characteristic information corresponding to each of the at least one first candidate operating state information is obtained. The timing model may include at least one of: a cyclic neural network (Recurrent Neural Network, RNN) based timing model, a Bi-cyclic neural network based timing model (Bi-directional Recurrent Neural Network, bi-RNN) and a transducer (i.e., converter) based timing model. The recurrent neural network model may include at least one of: long Short Term Memory network (LSTM) and gated loop unit (Gated Recurrent Unit, GRU), etc. The Bi-directional cyclic convolutional neural network model may include a Bi-directional long-short Term Memory network (Bi-directional Long Short-Term Memory, bi-LSTM), and the like. Further, the transducer-based timing model may include at least one of a base transducer and a deformation model for the transducer. The long-term memory network may include at least one of a basic long-term memory network and an improved network for the basic long-term memory network. The two-way long and short term memory network may include at least one of a basic two-way long and short term memory network and a modified network for the basic two-way long and short term memory network. The gating loop unit may include at least one of a basic gating loop unit and a modified network for the basic gating loop unit.

The reason for "extracting the time series feature of the first candidate operation state information to obtain the candidate operation state feature information" will be described below.

In the course of implementing the disclosed inventive concepts, time variability was found to affect both pipeline characteristic changes of the target device as well as periodic changes in environmental information.

Taking the example that the first candidate device control information includes the candidate cooling pump frequency and the candidate freezing pump frequency, the change in the pipe characteristics of the time-differential-influence target device is described. In theory, the cooling pump frequency and the cooling water flow rate of the refrigeration system should exhibit a unique linear relationship, and the freezing pump frequency and the freezing water flow rate of the refrigeration system should exhibit a unique linear relationship. However, by analyzing the relationship data between the cooling pump frequency and the cooling water flow rate, it was found that the cooling pump frequency and the cooling water flow rate exhibited a plurality of linear relationships. By analyzing the relationship data between the cryopump frequency and the flow of chilled water, it was found that the cryopump frequency and the flow of chilled water would exhibit a plurality of linear relationships. For example, the cooling pump frequency corresponding to each period is different for the same cooling water flow rate because the pipe characteristics of the target equipment change over time, for example, pipe network resistance changes due to at least one of pipe blockage, pipe aging, and the like will occur.

For periodic changes in environmental information, for example, the environmental information includes a dry bulb temperature and a wet bulb temperature, as the dry bulb temperature and the wet bulb temperature themselves may exhibit periodic changes over time. Taking the relationship data between the total secondary flow and wet bulb temperature as an example, the periodic variation of the environmental information will be described. For example, in the case of "year" as the study granularity, the secondary side total flow rate exhibited a positive correlation with wet bulb temperature. In the case of "day" as the study particle size, the secondary side total flow rate was periodic with wet bulb temperature.

Based on the analysis, since the first candidate operation state information includes the first candidate device control information and the current environment information, at least one first candidate state sub-information included in the first candidate operation state information has time sequence property, and accords with the time sequence characteristic, so that the first candidate operation state information can be subjected to time sequence characteristic extraction to obtain candidate operation state characteristic information.

According to an embodiment of the present disclosure, the at least one first candidate operating state information may include M. The mth first candidate operating state information may include T first candidate operating state sub-information. M may be an integer greater than or equal to 1. T may be an integer greater than 1. M may be an integer greater than or equal to 1 and less than or equal to M.

According to an embodiment of the present disclosure, performing time sequence feature extraction on at least one first candidate operation state information to obtain candidate operation state feature information corresponding to the at least one first candidate operation state information may include the following operations.

And responding to the T which is more than 1 and less than or equal to T, and determining forgetting characteristic information of the T first candidate running state sub-information, first updating characteristic information of the T first candidate running state sub-information and candidate state characteristic information of the T first candidate running state sub-information according to the T first candidate running state sub-information and the first hidden characteristic information of the T-1 first candidate running state sub-information. And determining the first state characteristic information of the t first candidate operation state sub-information according to the first state characteristic information of the t-1 first candidate operation state sub-information, the forgetting characteristic information of the t first candidate operation state sub-information, the first updated characteristic information of the t first candidate operation state sub-information and the candidate state characteristic information of the t first candidate operation state sub-information. And determining the first hidden characteristic information of the t first candidate running state sub-information according to the t first candidate running state sub-information, the first hidden characteristic information of the t-1 first candidate running state sub-information and the first state characteristic information of the t first candidate running state sub-information. And determining the mth candidate running state characteristic information according to the first hidden characteristic information of the first candidate running state sub-information of the mth.

According to an embodiment of the present disclosure, the mth candidate operating state characteristic information may characterize candidate operating state characteristic information of the mth first candidate operating state information. T may be an integer greater than or equal to 1 and less than or equal to T. U may be an integer greater than or equal to 1 and less than or equal to T. The values of M and T may be configured according to actual service requirements, and are not limited herein.

According to an embodiment of the present disclosure, the mth first candidate operation state information may include T first candidate operation state sub-information. For example, the mth first candidate operating state information may include the 1 st first candidate operating state sub-information, the 2 nd first candidate operating state sub-information, …, the T-th first candidate operating state sub-information, … …, the T-1 st first candidate operating state sub-information, and the T-th first candidate operating state sub-information. The T first candidate operating state sub-information is first candidate operating state sub-information having a timing relationship. The collection time of the t first candidate operation state sub-information is smaller than the collection time of the t-1 first candidate operation state sub-information. t.epsilon. {1,2, … …, T-1, T }.

According to an embodiment of the present disclosure, in case T > 1, in response to 1 < t+.t, for the mth first candidate operation state information of the M first candidate operation state information, determining forgetting feature information of the T first candidate operation state information, first update feature information of the T first candidate operation state information, and candidate state feature information of the T first candidate operation state information according to the T first candidate operation state information and first hidden feature information of the T-1 th first candidate operation state information, may include: and obtaining forgetting characteristic information of the t first candidate running state sub-information according to the first weight, the second weight, the t first candidate running state sub-information and the first hidden characteristic information of the t-1 first candidate running state sub-information. And obtaining first updated characteristic information of the t first candidate running state sub-information according to the third weight, the fourth weight, the t first candidate running state sub-information and the first hidden characteristic information of the t-1 first candidate running state sub-information. And obtaining candidate state characteristic information of the t first candidate running state sub-information according to the fifth weight, the sixth weight, the t first candidate running state sub-information and the first hidden characteristic information of the t-1 first candidate running state sub-information.

According to an embodiment of the present disclosure, obtaining forgetting feature information of the t first candidate operating state sub information according to the first weight, the second weight, the t first candidate operating state sub information and the first hidden feature information of the t-1 th first candidate operating state sub information may include: and obtaining first weighting characteristic information of the t first candidate running state sub-information according to the first weight and the t first candidate running state sub-information. And obtaining second weighted characteristic information of the t first candidate running state sub-information according to the second weight and the first hidden characteristic information of the t-1 first candidate running state sub-information. And obtaining forgetting characteristic information of the t first candidate running state sub-information according to the first weighted characteristic information and the second weighted characteristic information of the t first candidate running state sub-information.

According to an embodiment of the present disclosure, obtaining first updated feature information of the t first candidate operating state sub information according to the third weight, the fourth weight, the t first candidate operating state sub information and the first hidden feature information of the t-1 th first candidate operating state sub information may include: and obtaining third weighted characteristic information of the t lens according to the third weight and the t first candidate running state sub-information. And obtaining fourth weighted characteristic information of the t first candidate running state sub-information according to the fourth weight and the first hidden characteristic information of the t-1 first candidate running state sub-information. And obtaining first updated characteristic information of the t first candidate running state sub-information according to the third weighted characteristic information and the fourth weighted characteristic information of the t first candidate running state sub-information.

According to an embodiment of the present disclosure, obtaining candidate state feature information of the t first candidate operating state sub-information according to the fifth weight, the sixth weight, the t first candidate operating state sub-information, and the first hidden feature information of the t-1 first candidate operating state sub-information may include: and obtaining fifth weighted characteristic information of the t first candidate running state sub-information according to the fifth weight and the t first candidate running state sub-information. And obtaining sixth weighted characteristic information of the t first candidate running state sub-information according to the sixth weight and the first hidden characteristic information of the t-1 first candidate running state sub-information. And obtaining candidate state characteristic information of the t first candidate operation state sub-information according to the fifth weighted characteristic information and the sixth weighted characteristic information of the t first candidate operation state sub-information.

According to an embodiment of the present disclosure, determining the first state characteristic information of the t first candidate operating state sub-information according to the first state characteristic information of the t-1 th first candidate operating state sub-information, the forgetting characteristic information of the t first candidate operating state sub-information, the first updated characteristic information of the t first candidate operating state sub-information, and the candidate state characteristic information of the t first candidate operating state sub-information may include: and obtaining the first characteristic information of the t first candidate running state sub-information according to the first state characteristic information of the t-1 first candidate running state sub-information and the forgetting characteristic information of the t first candidate running state sub-information. And obtaining second characteristic information of the t first candidate running state sub-information according to the first updated characteristic information of the t first candidate running state sub-information and the candidate state characteristic information of the t first candidate running state sub-information. And obtaining the first state characteristic information of the t first candidate running state sub-information according to the first characteristic information and the second characteristic information of the t first candidate running state sub-information.

According to an embodiment of the present disclosure, determining the first hidden characteristic information of the t first candidate operating state sub information according to the t first candidate operating state sub information, the first hidden characteristic information of the t-1 th first candidate operating state sub information, and the first state characteristic information of the t first candidate operating state sub information may include: and obtaining output characteristic information of the t first candidate running state sub-information according to the t first candidate running state sub-information and the first hidden characteristic information of the t-1 first candidate running state sub-information. And obtaining first hidden characteristic information of the t first candidate running state sub-information according to the output characteristic information and the first state characteristic information of the t first candidate running state sub-information. Obtaining output feature information of the t first candidate running state sub-information according to the t first candidate running state sub-information and the first hidden feature information of the t-1 first candidate running state sub-information may include: and obtaining output characteristic information of the t first candidate running state sub-information according to the seventh weight, the eighth weight, the t first candidate running state sub-information and the first hidden characteristic information of the t-1 first candidate running state sub-information. For example, the seventh weighted feature information of the t first candidate operating state sub information may be obtained according to the seventh weight and the t first candidate operating state sub information. And obtaining eighth weighted characteristic information of the t first candidate running state sub-information according to the eighth weight and the first hidden characteristic information of the t-1 first candidate running state sub-information. And obtaining output characteristic information of the t first candidate running state sub-information according to the seventh weighted characteristic information and the eighth weighted characteristic information of the t first candidate running state sub-information.

According to the embodiment of the disclosure, in the case of t=1 or in the case of T > 1, in response to t=1, forgetting feature information of the 1 st first candidate operation state sub-information, first update feature information of the 1 st first candidate operation state sub-information and candidate state feature information of the 1 st first candidate operation state sub-information are obtained according to the 1 st first candidate operation state sub-information. And obtaining the first state characteristic information of the 1 st first candidate running state sub-information according to the forgetting characteristic information of the 1 st first candidate running state sub-information, the first updated characteristic information of the 1 st first candidate running state sub-information and the candidate state characteristic information of the 1 st first candidate running state sub-information. And obtaining first hidden characteristic information of the 1 st first candidate running state sub-information according to the 1 st first candidate running state sub-information and the first state characteristic information of the 1 st first candidate running state sub-information. And obtaining candidate running state characteristic information of the 1 st first candidate running state sub-information according to the first hidden characteristic information of the 1 st first candidate running state sub-information.

According to an embodiment of the present disclosure, determining the mth candidate operating state characteristic information according to the first hidden characteristic information of the mth first candidate operating state sub-information may include: the mth candidate operating state characteristic information may be determined according to the first hidden characteristic information of the mth first candidate operating state sub-information. For example, the first hidden feature information of the T-th first candidate operating state sub-information may be determined as the m-th candidate operating state feature information.

According to an embodiment of the present disclosure, the at least one first candidate operation state information may include P. The p-th first candidate operating state information may include Q first candidate operating state sub-information. P may be an integer greater than or equal to 1. Q may be an integer greater than 1. P may be an integer greater than or equal to 1 and less than or equal to P.

According to an embodiment of the present disclosure, performing time sequence feature extraction on at least one first candidate operation state information to obtain candidate operation state feature information corresponding to at least one first candidate operation state sub-information may include the following operations.

And responding to the Q which is more than 1 and less than or equal to Q, and determining second updated characteristic information of the Q first candidate running state sub-information and reset characteristic information of the Q first candidate running state sub-information according to the Q first candidate running state sub-information and the second hidden characteristic information of the Q-1 first candidate running state sub-information. And determining the second state characteristic information of the q first candidate operation state sub-information according to the q first candidate operation state sub-information, the second hidden characteristic information of the q-1 first candidate operation state sub-information and the reset characteristic information of the q first candidate operation state sub-information. And determining the second hidden characteristic information of the q first candidate running state sub-information according to the second updated characteristic information of the q first candidate running state sub-information, the second hidden characteristic information of the q-1 first candidate running state sub-information and the second state characteristic information of the q first candidate running state sub-information. And determining p candidate running state characteristic information according to the second hidden characteristic information of the V first candidate running state sub-information.

According to an embodiment of the present disclosure, the p-th candidate operating state characteristic information may characterize candidate operating state characteristic information of the p-th first candidate operating state information. Q may be an integer greater than or equal to 1 and less than or equal to Q. V may be an integer greater than or equal to 1 and less than or equal to Q.

According to an embodiment of the present disclosure, the p-th first candidate operation state information may include Q first candidate operation state sub-information. For example, the p-th first candidate operating state information may include the 1 st first candidate operating state sub-information, the 2 nd first candidate operating state sub-information, … …, the Q-th first candidate operating state sub-information, … …, the Q-1 st first candidate operating state sub-information, and the Q-th first candidate operating state sub-information. The Q first candidate operating state sub-information is first candidate operating state sub-information having a timing relationship. The collection time of the q-th first candidate operation state sub-information is smaller than the collection time of the q-1-th first candidate operation state sub-information. Q ε {1,2, … …, Q-1, Q }.

According to an embodiment of the present disclosure, in case Q > 1, in response to 1 < q+.q, for the P-th first candidate operating state information among the P-th first candidate operating state information, determining the second updated characteristic information of the Q-th first candidate operating state information and the reset characteristic information of the Q-th first candidate operating state information according to the Q-th first candidate operating state information and the second hidden characteristic information of the Q-1-th first candidate operating state information may include: and determining second updated characteristic information of the q first candidate running state sub-information according to the ninth weight, the q first candidate running state sub-information and the second hidden characteristic information of the q-1 first candidate running state sub-information. And determining reset characteristic information of the q first candidate running state sub-information according to the tenth weight, the q first candidate running state sub-information and the second hidden characteristic information of the q-1 first candidate running state sub-information.

According to an embodiment of the present disclosure, determining the second state characteristic information of the q-th first candidate operating state sub-information according to the q-th first candidate operating state sub-information, the second hidden characteristic information of the q-1-th first candidate operating state sub-information, and the reset characteristic information of the q-th first candidate operating state sub-information may include: and determining the second state characteristic information of the q first candidate operation state sub-information according to the eleventh weight, the q first candidate operation state sub-information, the second hidden characteristic information of the q-1 first candidate operation state sub-information and the reset characteristic information of the q first candidate operation state sub-information.

According to an embodiment of the present disclosure, determining the second hidden characteristic information of the q-th first candidate operating state sub-information according to the second updated characteristic information of the q-th first candidate operating state sub-information, the second hidden characteristic information of the q-1-th first candidate operating state sub-information, and the second state characteristic information of the q-th first candidate operating state sub-information may include: and obtaining first intermediate characteristic information of the q first candidate running state sub-information according to the second updated characteristic information of the q first candidate running state sub-information and the second hidden characteristic information of the q-1 first candidate running state sub-information. And obtaining second intermediate characteristic information of the q first candidate running state sub-information according to the second updated characteristic information of the q first candidate running state sub-information and the second state characteristic information of the q first candidate running state sub-information. And the first intermediate characteristic information and the second intermediate characteristic information of the q first candidate running state sub-information are obtained, so that the second hidden characteristic information of the q first candidate running state sub-information is obtained.

According to an embodiment of the present disclosure, in the case of q=1 or in the case of Q > 1, in response to q=1, the second update characteristic information of the 1 st first candidate operating state sub information and the reset characteristic information of the 1 st first candidate operating state sub information are determined from the 1 st first candidate operating state sub information. And determining second state characteristic information of the 1 st first candidate operation state sub-information according to the 1 st first candidate operation state sub-information and the reset characteristic information of the 1 st first candidate operation state sub-information. And obtaining first hidden characteristic information of the 1 st first candidate running state sub-information according to the 1 st first candidate running state sub-information and the first state characteristic information of the 1 st first candidate running state sub-information. And determining second hidden characteristic information of the 1 st first candidate running state sub-information according to the second updated characteristic information of the 1 st first candidate running state sub-information and the second state characteristic information of the 1 st first candidate running state sub-information.

According to an embodiment of the present disclosure, determining the p-th candidate operation state feature information according to the second hidden feature information of the V-th first candidate operation state sub-information may include: and determining p candidate running state characteristic information according to the second hidden characteristic information of the Q first candidate running state sub-information. For example, the second hidden characteristic information of the Q-th first candidate operating state sub-information may be determined as the p-th candidate operating state characteristic information.

And carrying out R layers of processing on the at least one first candidate operation state information based on the self-attention strategy to obtain candidate operation state characteristic information corresponding to the at least one first candidate operation state information.

According to embodiments of the present disclosure, R may be an integer greater than or equal to 1. The value of R may be configured according to actual service requirements, and is not limited herein. The first candidate operating state information may include at least one first candidate operating state sub-information. The at least one first candidate operating state sub-information may have a timing relationship.

According to embodiments of the present disclosure, a self-attention strategy may be used to achieve focusing important information with high weight, ignoring non-important information with low weight, and enabling information exchange with other information by sharing important information, thereby achieving the transfer of important information. In the embodiment of the disclosure, the self-attention strategy can extract the first candidate running state sub-information and the information between each first candidate running state sub-information so as to better complete the determination of the candidate running state characteristic information.

According to embodiments of the present disclosure, the first candidate run state sub-information may be used to determine a Key (i.e., key) matrix, a Value (i.e., value) matrix, and a Query (i.e., query) matrix. For example, the first candidate operating state sub-information may be used as a key matrix, a value matrix, and a query matrix. The key matrix, the values, and the query matrix may be matrices in an attention mechanism.

According to the embodiment of the disclosure, for the first candidate operation state information in the at least one first candidate operation state information, R levels of processing may be performed on a matrix set corresponding to the first candidate operation state information based on a self-attention policy, so as to obtain candidate operation state feature information corresponding to the first candidate operation state information. For example, the self-attention unit may be determined according to a self-attention policy. And carrying out R-level processing on the matrix set corresponding to the first candidate running state information by utilizing self-attention unit processing to obtain candidate running state characteristic information corresponding to the first candidate running state information. The matrix set may include matrix subsets each corresponding to at least one candidate operating state sub-information included in the first candidate operating state information. The matrix subset may include a key matrix, a value matrix, and a query matrix.

According to the embodiment of the disclosure, R layers of processing are performed on the first candidate running state information based on the self-attention strategy to obtain the candidate running state characteristic information corresponding to the first candidate running state information, and the self-attention strategy can extract information between the first candidate running state sub-information and each first candidate running state sub-information, so that the accuracy of the candidate running state characteristic information is improved.

According to an embodiment of the present disclosure, in a case where R is an integer greater than 1, performing R-level processing on at least one first candidate operation state information based on a self-attention policy, obtaining candidate operation state feature information corresponding to the at least one first candidate operation state information may include the following operations.

And responding to R which is more than 1 and less than or equal to R, and obtaining second intermediate candidate running state characteristic information of the R-1 level corresponding to the first candidate running state information according to the first intermediate candidate running state characteristic information of the R-1 level corresponding to the first candidate running state information. And obtaining first intermediate candidate running state characteristic information of the r-th level corresponding to the first candidate running state information according to the second intermediate candidate running state characteristic information of the r-th level corresponding to the first candidate running state information and the first intermediate candidate running state characteristic information of the r-1-th level corresponding to the first candidate running state information. And obtaining candidate running state characteristic information corresponding to the first candidate running state information according to the first intermediate candidate running state characteristic information of the S-level, which corresponds to the first candidate running state information.

According to embodiments of the present disclosure, R may be an integer greater than or equal to 1 and less than or equal to R. r.epsilon. {1,2, … …, R-1, R }. S may be an integer greater than 1 and less than or equal to R. The first intermediate candidate operating state characteristic information is used to determine a query matrix, a key matrix, and a value matrix.

According to the embodiment of the disclosure, in the case that R is more than 1, responding to 1 < r.ltoreq.R, aiming at first candidate operation state information in at least one first candidate operation state information, processing first intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of an R-1 level based on a self-attention strategy, and obtaining second intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the R level. The second intermediate candidate operating state characteristic information of the r-th hierarchy corresponding to the first candidate operating state information may be used as a key matrix set, a value matrix set, and a query matrix set of the r+1-th hierarchy. And carrying out fusion processing on the second intermediate candidate running state characteristic information of the r-th level corresponding to the first candidate running state information and the first intermediate candidate running state characteristic information of the r-1-th level corresponding to the first candidate running state information to obtain sixteenth intermediate candidate running state characteristic information of the r-th level corresponding to the first candidate running state information. And obtaining first intermediate candidate running state characteristic information of the r level corresponding to the first candidate running state information according to the sixteenth intermediate candidate running state characteristic information of the r level corresponding to the first candidate running state information.

According to an embodiment of the present disclosure, obtaining first intermediate candidate operation state feature information of the r-th hierarchy corresponding to the first candidate operation state information according to sixteenth intermediate candidate operation state feature information of the r-th hierarchy corresponding to the first candidate operation state information may include: and processing sixteenth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the r-th level based on the first multi-layer perceptron strategy to obtain seventeenth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the r-th level. And obtaining first intermediate candidate running state characteristic information of the r-th level corresponding to the first candidate running state information according to the seventeenth intermediate candidate running state characteristic information of the r-th level corresponding to the first candidate running state information. For example, the seventeenth intermediate candidate operation state characteristic information of the r-th level and the first candidate operation state information may be normalized, so as to obtain the first intermediate candidate operation state characteristic information of the r-th level and corresponding to the first candidate operation state information. Normalization (i.e., normalization) may include one of: batch normalization (Batch Normalization, BN) and layer normalization (Lay Normalization, LN). For example, batch normalization processing may be performed on seventeenth intermediate candidate operation state feature information of the r-th level and the first candidate operation state information, so as to obtain first intermediate candidate operation state feature information of the r-th level and corresponding to the first candidate operation state information.

According to the embodiment of the disclosure, the first intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the r-1 th level is processed based on the self-attention strategy, and the second intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the r-1 th level is obtained. And obtaining the eighteenth intermediate candidate running state characteristic information of the r-1 level and the first candidate running state information according to the seventeenth intermediate candidate running state characteristic information of the r-1 level and the first candidate running state information. For example, the seventeenth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information in the r-1 level is subjected to standardization processing, so that the eighteenth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information in the r level is obtained. And processing the eighteenth intermediate candidate running state characteristic information of the nth hierarchy corresponding to the first candidate running state information based on the self-attention strategy to obtain first intermediate candidate running state characteristic information of the nth hierarchy corresponding to the first candidate running state information.

According to an embodiment of the present disclosure, in the case of r=1 or in the case of R > 1, for first candidate operation state information of at least one first candidate operation state information, processing the first candidate operation state information of the 1 st level based on the self-attention policy, to obtain second intermediate candidate operation state feature information of the 1 st level corresponding to the first candidate operation state information, in response to r=1. And carrying out fusion processing on the second intermediate candidate running state characteristic information corresponding to the first candidate running state information of the 1 st level and the first candidate running state information of the 1 st level to obtain sixteenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the 1 st level. And obtaining first intermediate candidate running state characteristic information corresponding to the first candidate running state information according to the sixteenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the 1 st level.

According to the embodiment of the disclosure, candidate operation state characteristic information corresponding to the first candidate operation state information is obtained according to the first intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the S-level. For example, the first intermediate candidate operation state characteristic information corresponding to the first candidate operation state information may be obtained according to the first intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the R-th hierarchy. The fusion may include one of the following: adding and splicing.

And processing the at least one first candidate operation state information based on the sparse self-attention strategy to obtain third intermediate candidate operation state characteristic information corresponding to the at least one first candidate operation state information. And processing the third intermediate candidate operation state characteristic information corresponding to the at least one first candidate operation state information based on the self-focusing distillation strategy to obtain fourth intermediate candidate operation state characteristic information corresponding to the at least one first candidate operation state information. And obtaining candidate operation state characteristic information corresponding to the at least one first candidate operation state information according to the at least one first candidate operation state information and fourth intermediate candidate operation state characteristic information corresponding to the at least one first candidate operation state information.

According to embodiments of the present disclosure, a sparse self-attention policy may refer to a self-attention policy for determining a sparse query matrix. The sparse query matrix may be determined based on the relative entropy between the key matrix and the query matrix. The first candidate operating state information may be used to determine a key matrix and a query matrix. A self-attention distillation strategy may refer to a strategy for achieving a higher weighting of features with dominant features and generating a focused self-attention feature map.

According to an embodiment of the present disclosure, the first candidate operating state information may include at least one first candidate operating state sub-information. The at least one first candidate operating state sub-information may have a timing relationship. The first candidate operating state sub-information may be used to determine a key matrix, a value matrix, and a query matrix. For first candidate operation state information in at least one first candidate operation state information, for first candidate operation state sub-information in at least one first candidate operation state sub-information included in the first candidate operation state information, determining at least one target key matrix corresponding to the first candidate operation state sub-information from at least one key matrix corresponding to the first candidate operation state sub-information. And determining the relative entropy between the query matrix and each of the at least one target key matrix according to the query matrix and the at least one target key matrix corresponding to the first candidate running state sub-information. And obtaining sparsity values corresponding to the query matrix according to the relative entropy between the query matrix and each of the at least one target key matrix. At least one target query matrix is determined from the query matrices corresponding to each of the at least one first candidate operating state sub-information based on the sparsity values of the query matrices corresponding to each of the at least one first candidate operating state sub-information. An attention value corresponding to the first candidate run state sub-information is determined from the at least one target query matrix, the at least one target key matrix corresponding to the at least one target query matrix, and the at least one value matrix corresponding to the at least one target query matrix. And obtaining third intermediate candidate operation state characteristic information corresponding to the first candidate operation state information according to the attention value corresponding to the at least one first candidate operation state sub-information.

According to the embodiment of the disclosure, for first candidate operation state information in at least one first candidate operation state information, nineteenth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information is obtained according to third intermediate candidate operation state characteristic information corresponding to the first candidate operation state information. And obtaining twenty-first intermediate candidate operation state characteristic information corresponding to the first candidate operation state information according to nineteenth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information. Pooling the twentieth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information to obtain fourth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information.

According to the embodiment of the disclosure, the calculation complexity and the space complexity of calculating the self-attention are reduced by processing the at least one first candidate operation state information based on the sparse self-attention strategy to obtain the third intermediate candidate operation state characteristic information corresponding to the at least one first candidate operation state information. In addition, by processing the third intermediate candidate operating state characteristic information corresponding to the at least one first candidate operating state information based on the self-care distillation strategy, fourth intermediate candidate operating state characteristic information corresponding to the at least one first candidate operating state information is obtained, reducing the overall spatial complexity. Thereby, resource consumption is reduced.

According to an embodiment of the present disclosure, in a case where J may be an integer greater than 1, performing time-series feature extraction on the at least one first candidate operation state information, obtaining candidate operation state feature information corresponding to the at least one first candidate operation state information may include the following operations.

Responding to 1 < j.ltoreq.J, and processing the first seasonal characteristic information corresponding to the first candidate running state information of the J-1 th level based on the autocorrelation strategy to obtain the autocorrelation characteristic information corresponding to the first candidate running state information of the J-1 th level. And obtaining fifth intermediate candidate running state characteristic information of the j-th level, which corresponds to the first candidate running state information, according to the autocorrelation characteristic information of the j-th level, which corresponds to the first candidate running state information, and the first seasonal characteristic information of the j-1-th level, which corresponds to the first candidate running state information. And processing fifth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the j-th level based on the first sequence decomposition strategy to obtain second seasonal characteristic information corresponding to the first candidate running state information of the j-th level. And obtaining first seasonal characteristic information of the j-th level corresponding to the first candidate running state information according to the second seasonal characteristic information of the j-th level corresponding to the first candidate running state information. And obtaining candidate running state characteristic information corresponding to the first candidate running state information according to the first seasonal characteristic information corresponding to the first candidate running state information of the K-th level.

According to embodiments of the present disclosure, J may be an integer greater than or equal to 1 and less than or equal to J. K may be an integer greater than or equal to 1 and less than or equal to J.

According to an embodiment of the present disclosure, the first candidate operating state information may include at least one first candidate operating state sub-information. The at least one first candidate operating state sub-information has a timing relationship.

According to embodiments of the present disclosure, an autocorrelation strategy may be used to implement dependency discovery (i.e., period-based Dependencies) and delay information aggregation (i.e., time Delav Aggregation) at the sequence level. The autocorrelation strategy is used to determine a timing periodicity between at least one first candidate operating state sub-information included in the first candidate operating state information. The autocorrelation strategies may include a cycle-based discovery-dependent strategy and a latency information aggregation-based strategy. The first sequence decomposition policy may refer to a seasonal policy for determining the first candidate operating state information.

According to an embodiment of the present disclosure, in case of J > 1, in response to 1 < j+.j, for first candidate operation state information among at least one first candidate operation state information, at least one first autocorrelation coefficient corresponding to the first candidate operation state information of the J-1 th hierarchy is determined according to first seasonal feature information corresponding to the first candidate operation state information of the J-1 th hierarchy and other first seasonal feature information corresponding to the first candidate operation state information of the J-1 th hierarchy. And determining at least one target autocorrelation coefficient of the j-1 th level corresponding to the first candidate running state information according to at least one first autocorrelation coefficient of the j-1 th level corresponding to the first candidate running state information. And obtaining the auto-correlation characteristic information of the j-th level corresponding to the first candidate running state information according to at least one other first seasonal characteristic information of the j-1-th level corresponding to the first candidate running state information and at least one target auto-correlation coefficient corresponding to the first candidate running state information. Other first seasonal feature information corresponding to the first candidate operation state information may be obtained by time-delaying the first seasonal feature information corresponding to the first candidate operation state information.

According to the embodiment of the disclosure, the auto-correlation characteristic information corresponding to the first candidate operation state information of the jth level and the first seasonal characteristic information corresponding to the first candidate operation state information of the jth-1 level can be fused to obtain fifth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the jth level. And processing fifth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the j-th level based on the first pooling strategy to obtain first trend characteristic information corresponding to the first candidate running state information of the j-th level. And obtaining second seasonal feature information of the j-th level corresponding to the first candidate running state information according to the fifth intermediate candidate running state feature information and the first trend feature information of the j-th level corresponding to the first candidate running state information.

According to an embodiment of the present disclosure, in a case where j=1 or in a case where J > 1, for first candidate operation state information among at least one first candidate operation state information, at least one first autocorrelation coefficient of the 1 st stage corresponding to the first candidate operation state information is determined from the first candidate operation state information of the 1 st stage and other candidate operation state information of the 1 st stage in response to j=1. And determining at least one target autocorrelation coefficient corresponding to the first candidate running state information of the 1 st level according to the at least one first autocorrelation coefficient corresponding to the first candidate running state information of the 1 st level. And obtaining the autocorrelation characteristic information of the 1 st level corresponding to the first candidate running state information according to at least one target autocorrelation coefficient corresponding to the first candidate running state information of the 1 st level and other candidate running state information. The other candidate running state information may be obtained by performing time delay on the first seasonal feature information corresponding to the first candidate running state information.

According to the embodiment of the disclosure, the autocorrelation characteristic information corresponding to the first candidate running state information in the 1 st level and the first candidate running state information in the 1 st level can be fused to obtain fifth intermediate candidate running state characteristic information corresponding to the first candidate running state information in the 1 st level. And processing fifth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the 1 st level based on the first pooling strategy to obtain first trend characteristic information corresponding to the first candidate running state information of the 1 st level. And obtaining second seasonal characteristic information corresponding to the first candidate running state information of the 1 st level according to the fifth intermediate candidate running state characteristic information and the first trend characteristic information corresponding to the first candidate running state information of the 1 st level. And obtaining first seasonal characteristic information corresponding to the first candidate running state information of the 1 st level according to the second seasonal characteristic information corresponding to the first candidate running state information of the 1 st level.

According to the embodiment of the disclosure, candidate operation state characteristic information corresponding to the first candidate operation state information is obtained according to the first seasonal characteristic information corresponding to the first candidate operation state information of the K-th level. For example, candidate operation state feature information corresponding to the first candidate operation state information may be obtained according to first seasonal feature information corresponding to the first candidate operation state information of the J-th hierarchy.

According to the embodiment of the disclosure, the sequence level connection and the complexity are realized based on the autocorrelation strategy, so that the bottleneck of information utilization is broken.

According to an embodiment of the present disclosure, obtaining the first seasonal feature information corresponding to the first candidate operation state information of the jth level according to the second seasonal feature information corresponding to the first candidate operation state information of the jth level may include the following operations.

And obtaining sixth intermediate candidate running state characteristic information of the j-th level, which corresponds to the first candidate running state information, according to the second seasonal characteristic information of the j-th level, which corresponds to the first candidate running state information. And obtaining seventh intermediate candidate running state characteristic information of the j-th level corresponding to the first candidate running state information according to the second seasonal characteristic information of the j-th level corresponding to the first candidate running state information and the sixth intermediate candidate running state characteristic information of the j-th level corresponding to the first candidate running state information. And processing the seventh intermediate candidate running state characteristic information corresponding to the first candidate running state information of the j-th level based on a second sequence decomposition strategy to obtain first seasonal characteristic information corresponding to the first candidate running state information of the j-th level.

According to the embodiment of the disclosure, in the case of j=1 or in the case of J > 1, in response to=1, processing the second seasonal feature information corresponding to the first candidate operation state information of the J-th level based on the second multi-layer perceptron policy, to obtain sixth intermediate candidate operation state feature information corresponding to the first candidate operation state information of the J-th level. And carrying out fusion processing on the second seasonal characteristic information corresponding to the first candidate running state information of the j-th level and the sixth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the j-th level to obtain seventh intermediate candidate running state characteristic information corresponding to the first candidate running state information of the j-th level. And processing the seventh intermediate candidate running state characteristic information corresponding to the first candidate running state information of the j-th level based on the second pooling strategy to obtain second trend characteristic information corresponding to the first candidate running state information of the j-th level. And obtaining first seasonal feature information of the j-th level corresponding to the first candidate running state information according to the seventh intermediate candidate running state feature information and the second trend feature information of the j-th level corresponding to the first candidate running state information.

According to the embodiment of the disclosure, the efficiency of data retrieval and extraction is improved by considering the similarity between elements based on the local sensitive hash strategy, and the time effect is reduced.

According to an embodiment of the present disclosure, in a case where W is an integer greater than 1, performing time-series feature extraction on at least one first candidate operation state information to obtain candidate operation state feature information corresponding to the at least one first candidate operation state information may include the following operations.

And responding to W is more than 1 and less than or equal to W, and obtaining ninth intermediate candidate running state characteristic information of the W-1 level, which corresponds to the first candidate running state information, on the basis of the local sensitive Hash attention strategy for the eighth intermediate candidate running state characteristic information of the W-1 level. And obtaining tenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level according to the ninth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level and the eighth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-1-th level. And processing tenth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the w-th level based on the first reversible residual strategy to obtain tenth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the w-th level. And obtaining eighth intermediate candidate running state characteristic information corresponding to the first candidate running state information according to the tenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level. And obtaining candidate running state characteristic information corresponding to the first candidate running state information according to the eighth intermediate running state characteristic information of the O level, which corresponds to the first candidate running state information.

According to embodiments of the present disclosure, W may be an integer greater than or equal to 1 and less than or equal to W, and O may be an integer greater than or equal to 1 and less than or equal to W.

According to an embodiment of the present disclosure, the first candidate operating state information may include at least one first candidate operating state sub-information. The at least one first candidate operating state sub-information has a timing relationship. The locality sensitive hash attention policy may refer to a self-attention policy based on locality sensitive hashes. The locality sensitive hash attention policy may comprise a multi-round locality sensitive hash attention policy.

According to the embodiment of the disclosure, under the condition that W is greater than 1, in response to 1 < w.ltoreq.L, the ninth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the W-th level and the eighth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the W-th level can be subjected to fusion processing, so as to obtain tenth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the W-th level.

According to the embodiment of the disclosure, in the case of w=1 or in the case of W > 1, in response to w=1, the first candidate operation state information of the W-th level may be obtained based on the locality sensitive hash attention policy, so as to obtain the ninth intermediate candidate operation state feature information of the W-th level, which corresponds to the first candidate operation state information. And carrying out fusion processing on the ninth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w th level and the first candidate running state information of the w th level to obtain tenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w th level.

According to the embodiment of the disclosure, candidate operation state characteristic information corresponding to the first candidate operation state information is obtained according to eighth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the O level. For example, candidate operation state feature information corresponding to the first candidate operation state information may be obtained according to eighth intermediate candidate operation state feature information corresponding to the first candidate operation state information of the W-th hierarchy.

According to an embodiment of the present disclosure, obtaining the eighth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the w-th hierarchy according to the tenth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the w-th hierarchy may include the following operations.

And obtaining eleventh intermediate candidate running state characteristic information corresponding to the first candidate running state information according to the tenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level. And obtaining twelfth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level according to the tenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level and the eleventh intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level. And processing twelfth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level based on the second reversible residual strategy to obtain eighth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level.

According to the embodiment of the disclosure, in the case of w=1 or in the case of W > 1, in response to w=1, tenth intermediate candidate operation state feature information corresponding to the first candidate operation state information of the W-th hierarchy may be processed based on the third multi-layer perceptron policy, to obtain eleventh intermediate candidate operation state feature information corresponding to the first candidate operation state information of the W-th hierarchy. And carrying out fusion processing on tenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level and eleventh intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level to obtain twelfth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level.

According to an embodiment of the present disclosure, in a case where I is an integer greater than 1, performing timing feature extraction on at least one first candidate operation state information to obtain candidate operation state feature information corresponding to the at least one first candidate operation state information may include the following operations.

And responding to the I which is more than 1 and less than or equal to I, and processing third seasonal characteristic information corresponding to the first candidate running state information of the ith hierarchy based on a frequency domain enhancement strategy to obtain frequency domain characteristic information corresponding to the first candidate running state information of the ith hierarchy. And obtaining thirteenth intermediate candidate running state characteristic information of the ith level, which corresponds to the first candidate running state information, according to the frequency domain characteristic information of the ith level, which corresponds to the first candidate running state information, and the third seasonal characteristic information of the ith-1 level, which corresponds to the first candidate running state information. And processing thirteenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the ith hierarchy based on the first mixed expert decomposition strategy to obtain fourth section characteristic information corresponding to the first candidate running state information of the ith hierarchy. And obtaining third seasonal characteristic information of the ith level corresponding to the first candidate running state information according to the fourth seasonal characteristic information of the ith level corresponding to the first candidate running state information. And obtaining candidate running state characteristic information corresponding to the first candidate running state information according to third seasonal characteristic information of the G level, which corresponds to the first candidate running state information.

According to embodiments of the present disclosure, I may be an integer greater than or equal to 1 and less than or equal to I. G may be an integer greater than or equal to 1 and less than or equal to I.

According to an embodiment of the present disclosure, the first candidate operating state information may include at least one first candidate operating state sub-information. The at least one first candidate operating state sub-information has a timing relationship. Frequency enhancement strategies may be used to implement self-attention mechanisms. The frequency enhancement strategy may include a fourier transform-based frequency enhancement strategy and a wavelet transform-based frequency enhancement strategy. The first hybrid expert decomposition strategy may be used to implement frequency-enhanced decomposition of seasonal characteristic information to better capture global characteristics of the time series.

According to the embodiment of the disclosure, under the condition that I is more than 1, in response to 1 < I being less than or equal to I, the frequency domain characteristic information corresponding to the first candidate running state information of the ith level and the third seasonal characteristic information corresponding to the first candidate running state information of the ith-1 level can be subjected to fusion processing, so that thirteenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the ith level is obtained.

According to the embodiment of the disclosure, in the case of i=1 or in the case of I > 1, in response to i=1, the first candidate operation state information of the 1 st level is processed based on the frequency domain enhancement policy, so as to obtain frequency domain feature information of the 1 st level corresponding to the first candidate operation state information. And obtaining thirteenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the 1 st level according to the frequency domain characteristic information corresponding to the first candidate running state information of the 1 st level and the first candidate running state information of the 1 st level. And processing thirteenth intermediate candidate running state characteristic information corresponding to the first candidate running state information on the 1 st level based on the first mixed expert decomposition strategy to obtain fourth section characteristic information corresponding to the first candidate running state information on the 1 st level. And obtaining third seasonal characteristic information corresponding to the first candidate running state information of the 1 st level according to the fourth seasonal characteristic information corresponding to the first candidate running state information of the 1 st level.

According to the embodiment of the disclosure, candidate operation state characteristic information corresponding to the first candidate operation state information is obtained according to third seasonal characteristic information corresponding to the first candidate operation state information of the G level. For example, candidate operation state feature information corresponding to the first candidate operation state information may be obtained according to third seasonal feature information corresponding to the first candidate operation state information of the I-th hierarchy.

According to an embodiment of the present disclosure, obtaining third seasonal feature information corresponding to the first candidate operation state information of the ith hierarchy according to fourth seasonal feature information corresponding to the first candidate operation state information of the ith hierarchy may include the following operations.

And obtaining fourteenth intermediate candidate running state characteristic information corresponding to the first candidate running state information according to the fourth joint characteristic information corresponding to the first candidate running state information of the ith hierarchy. And obtaining fifteenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the ith hierarchy according to the fourth joint characteristic information corresponding to the first candidate running state information of the ith hierarchy and the fourteenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the ith hierarchy. And processing fifteenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the ith hierarchy based on a second mixed expert decomposition strategy to obtain third seasonal characteristic information corresponding to the first candidate running state information of the ith hierarchy.

According to the embodiment of the disclosure, in the case of i=1 or in the case of I > 1, in response to i=1, fourth node feature information corresponding to the first candidate operation state information of the ith hierarchy may be processed based on the fourth multi-layer perceptron policy, to obtain fourteenth intermediate candidate operation state feature information corresponding to the first candidate operation state information of the ith hierarchy. And processing fourth-season node characteristic information corresponding to the first candidate running state information of the ith hierarchy and fourteenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the ith hierarchy to obtain fifteenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the ith hierarchy.

According to an embodiment of the present disclosure, determining nth fusion information corresponding to at least one first candidate operating state information of a target system according to candidate operating state feature information, nth-1 fusion information, and nth-1 candidate device energy consumption influence information may include the following operations.

And carrying out fusion processing on the candidate operation state characteristic information, the n-1 fusion information and the n-1 candidate equipment energy consumption influence information corresponding to at least one first candidate operation state information of the target system to obtain the n fusion information corresponding to the at least one first candidate operation state information.

According to the embodiment of the disclosure, for the first candidate operation state information in the at least one first candidate operation state information, the candidate operation state feature information, the n-1 th fusion information and the n-1 th candidate equipment energy consumption influence information corresponding to the first candidate operation state information can be fused to obtain the n-th fusion information corresponding to the first candidate operation state information.

According to an embodiment of the present disclosure, obtaining the nth candidate device energy consumption influence information corresponding to the at least one first candidate operation state information according to the nth fusion information corresponding to the at least one first candidate operation state information may include the following operations.

And processing the nth fusion information corresponding to the at least one first candidate operation state information based on the full connection strategy to obtain nth candidate equipment energy consumption influence information corresponding to the at least one first candidate operation state information.

According to embodiments of the present disclosure, the full connection layer may be determined based on a full connection policy. And processing the nth fusion information corresponding to each of the at least one first candidate operation state information by using the full connection layer to obtain nth candidate equipment energy consumption influence information corresponding to each of the at least one first candidate operation state information.

According to the embodiment of the disclosure, the energy consumption influence information determination model may be utilized to process the at least one first candidate operation state information, so as to obtain N levels of candidate device energy consumption influence information corresponding to each of the at least one first candidate operation state information.

According to an embodiment of the present disclosure, the energy consumption influence information determination model may be obtained by training a predetermined model using the sample operation state information. The predetermined model may include a gating loop unit and at least one fully connected layer.

According to the embodiment of the disclosure, for the first candidate operation state information in the at least one first candidate operation state information, the first candidate operation state information may be processed by using the gating loop unit to obtain candidate operation state feature information. And processing the candidate running state characteristic information by using at least one full connection layer to obtain the 1 st candidate equipment energy consumption influence information. And under the condition that N is more than 1, responding to the condition that 1 < N is less than or equal to N, and determining nth fusion information corresponding to the first candidate operation state information according to the candidate operation state characteristic information, the nth-1 fusion information and the nth-1 candidate equipment energy consumption influence information corresponding to the first candidate operation state information. And processing the nth fusion information by using at least one full connection layer to obtain the nth candidate equipment energy consumption influence information.

The following describes the application effect of the energy consumption influence information determination model by taking a refrigeration system based on a data center as an example.

The refrigeration system may include four refrigeration subsystems. The refrigeration subsystem may include a refrigeration unit, a cooling pump, a cooling tower, a primary pump, a plate heat exchanger, four secondary pumps, and a cold storage tank for use as a backup cold source. And respectively taking the refrigeration subsystem and the secondary side as a whole, and training a preset model by utilizing sample running state information to obtain an energy consumption influence information determination model. And the effectiveness of the energy consumption influence information determination model is verified by comparing the energy consumption influence information determination model with the effects of other models.

The predetermined model may include 3 full connection layers and 1 gating loop unit. The gated loop unit may include 2 layers. The fully connected layer may comprise 3 layers. The loss function used to train the predetermined model may be configured according to actual business requirements and is not limited herein. For example, the Loss function used to train the predetermined model may include SmoothL1Loss. Table 1 schematically shows the effect comparison of the energy consumption influence information determination model with other models. The evaluation metrics used to evaluate the model effect may include MSE (Mean Square Error ), MAE (Mean Absolute Error, mean absolute error), MAPE (Mean Absolute Percentage Error ), and RMSE (Root Mean Square Error, mean error).

TABLE 1

The information on the impact of energy consumption is identified by the Infer-GRU in Table 1. The FCN characterizes a fully connected network (Full Connected Network). As can be seen from Table 1, each of the evaluation criteria for the Infer-GRU is superior to the FCN and GRU. The secondary side includes a prediction of the total flow of the circuit, and thus the error of each evaluation index is relatively large.

In summary, the Infer-GRU provided by the embodiment of the disclosure has the capturing capability of the cyclic neural network on the time sequence relationship, the fitting capability of the deep learning model on the high-dimensional characteristics, and the causal relationship between the measurement points of the nested target system equipment, and can be expanded in various application scenes, for example, the application scenes of various refrigeration systems.

According to an embodiment of the present disclosure, the first current selection information may include current device energy consumption information. The candidate device energy consumption impact information for the at least one hierarchy may include candidate device energy consumption information and candidate demand information.

According to an embodiment of the present disclosure, operation S230 may include the following operations.

At least one second candidate operating state information is determined from the at least one first candidate operating state information based on the candidate demand information corresponding to the at least one first candidate operating state information. And determining target equipment control information meeting a preset energy saving condition from the current equipment control information of the current operation state information and the first candidate equipment control information of the at least one second candidate operation state information according to the current equipment energy consumption information corresponding to the current operation state information and the candidate equipment energy consumption information corresponding to the at least one second candidate operation state information.

According to an embodiment of the present disclosure, the candidate demand information may refer to a target amount generated in the case where the target system operation is controlled according to the first candidate device control information. The target amount may include at least one of: cold and heat. The candidate demand information may be characterized by a target amount. The candidate device energy consumption information may be characterized by candidate device energy consumption. The current device energy consumption information may be the current device energy consumption.

According to an embodiment of the present disclosure, for first candidate operation state information of at least one first candidate operation state information, in a case where it is determined that a target amount corresponding to the first candidate operation state information is greater than or equal to a predetermined amount, the first candidate operation state information may be determined as second candidate operation state information. The predetermined amount may be configured according to actual service requirements, and is not limited herein.

According to an embodiment of the present disclosure, a minimum device energy consumption is determined from the current device energy consumption and the at least one candidate device energy consumption. And determining the device control information corresponding to the minimum device energy consumption as target device control information. In the case where the minimum device power consumption is the current device power consumption, the current device control information corresponding to the current device power consumption information may be determined as the target device control information. In the case where the minimum device power consumption is the candidate device power consumption, the first candidate device control information corresponding to the candidate device power consumption information may be determined as the target device control information.

According to the embodiment of the disclosure, since the target device control information is determined from the current device control information of the current operation state information and the first candidate device control information of the at least one first candidate operation state information based on the current device energy consumption information and the candidate device energy consumption information and based on the candidate demand amount information, it is realized that the optimal device control information, i.e., the target device control information, is determined from a plurality of reliable device control information, and the reliability of the target device control information is improved.

According to an embodiment of the present disclosure, the first current selection information may further include current device operation mode information. The first candidate selection information may also include candidate device operating mode information.

According to an embodiment of the present disclosure, determining target device control information satisfying a predetermined power saving condition from current device control information of current operation state information and first candidate device control information of at least one second candidate operation state information according to current device energy consumption information corresponding to the current operation state information and candidate device energy consumption information corresponding to the at least one second candidate operation state information may include the following operations.

At least one third candidate operation state information is determined from the at least one second candidate operation state information according to the current device operation mode information corresponding to the current operation state information and the candidate device operation mode information corresponding to the at least one second candidate operation state information. And determining target equipment control information meeting a preset energy saving condition from the current equipment control information of the current operation state information and the first candidate equipment control information of the at least one third candidate operation state information according to the current equipment energy consumption information corresponding to the current operation state information and the candidate equipment energy consumption information corresponding to the at least one third candidate operation state information.

According to an embodiment of the present disclosure, the candidate device operation mode information may refer to device operation mode information of the target device corresponding to the first candidate device control information. The current device operation mode information may refer to device operation mode information of the target device corresponding to the current period.

According to the embodiment of the disclosure, the similarity between the current device operation mode information and the candidate device operation mode information corresponding to the at least one second candidate operation state information can be determined, and at least one fourth similarity is obtained. The fourth degree of similarity may characterize a degree of similarity between the current device operating mode information and the candidate device operating mode information. If the current device operation mode information and the candidate device operation mode information match, the similarity between the current device operation mode information and the candidate device operation mode information may be 1. If the current device operation mode information and the candidate device operation model information do not match, the similarity between the current device operation mode information and the candidate device operation mode information may be 0. The relationship between the fourth similarity and the degree of similarity may be configured according to the actual service requirement, which is not limited herein. And sequencing the at least one second candidate operation state information according to the at least one fourth similarity to obtain third sequencing information. And determining at least one third candidate operation state information from the at least one second candidate operation state information according to the third ordering information. Alternatively, at least one third candidate operating state information may be determined from the at least one second candidate operating state information according to a fourth predetermined similarity threshold and the at least one fourth similarity. The fourth predetermined similarity threshold may be configured according to actual service requirements, and is not limited herein. The fourth similarity may be a number greater than or equal to 0 and less than or equal to 1.

For example, the greater the fourth degree of similarity, the greater the degree of similarity. The second candidate operation state information corresponding to each of the at least one fourth similarity may be sorted in order of the fourth similarity from the higher order to the lower order, and a third predetermined number of second candidate operation state information in the front of the sorting may be determined from the at least one second candidate operation state information as the at least one third candidate operation state information. Alternatively, for the second candidate operation state information of the at least one second candidate operation state information, in a case where it is determined that the fourth similarity corresponding to the second candidate operation state information is greater than or equal to the fourth predetermined similarity threshold value, the second candidate operation state information may be determined as the third candidate operation state information. The third predetermined number may be configured according to actual service requirements, and is not limited herein.

The target device control information is adjusted in response to detecting that the current environmental information meets at least one of a predetermined environmental condition and the target device control information meets a predetermined control condition.

According to an embodiment of the present disclosure, the current context information may include at least one of: current temperature and current humidity. The current context information may be characterized by a current context value. The predetermined environmental condition may be determined based on a predetermined environmental threshold. The target device control information may include target device control dimension information of at least one dimension. The target device control dimension information may be characterized by a target device control dimension value. The predetermined control condition may be determined from predetermined device control dimension thresholds that each correspond to at least one target device control dimension. The predetermined environmental threshold and the predetermined device control dimension threshold may be configured according to actual traffic demands, and are not limited herein.

According to embodiments of the present disclosure, the target device control information may be adjusted in response to detecting that the current environmental value is greater than or equal to a predetermined environmental threshold and that the at least one target device control dimension value is less than a predetermined device control dimension threshold corresponding to the at least one target device control dimension value. Alternatively, the target device control information may be adjusted in response to detecting that the current environment value is less than the predetermined environment threshold and the at least one target device control dimension value is greater than or equal to a predetermined device control dimension threshold corresponding to the at least one target device control dimension value. Alternatively, the target device control information may be adjusted in response to detecting that the current environment value is greater than or equal to a predetermined environment threshold and that the at least one target device control dimension value is greater than or equal to a predetermined device control dimension threshold corresponding to the at least one target device control dimension value.

According to an embodiment of the present disclosure, it may be detected whether or not there is at least one of the current environmental information satisfying the predetermined environmental condition and the target device control information satisfying the predetermined control condition at a predetermined period. The target device control information is adjusted in response to detecting that the current environmental information satisfies at least one of a predetermined environmental condition and the target device control information satisfies the predetermined control condition. The predetermined period may be configured according to actual service requirements, and is not limited herein. For example, the predetermined period may be an interval of a predetermined length of time.

According to an embodiment of the present disclosure, the target device control information may be determined as the first device control information to be adjusted in response to detecting that the current environment value is greater than or equal to the predetermined environment threshold and that the at least one target device control dimension value is both less than the predetermined device control dimension threshold corresponding to the at least one target device control dimension value. And adjusting the first device control information to be adjusted in response to the current environmental change value being greater than or equal to the predetermined environmental change threshold. And in response to the current environmental change value being less than the predetermined environmental change threshold, maintaining the first device to be adjusted control information unchanged.

According to embodiments of the present disclosure, the target device control information may be determined as the second device control information to be adjusted in response to detecting that the current environment value is less than the predetermined environment threshold and the at least one target device control dimension value is greater than or equal to a predetermined device control dimension threshold corresponding to the at least one target device control dimension value. And adjusting the second device control information to be adjusted in response to the at least one target device control dimension change value being greater than or equal to a predetermined device control dimension change threshold. And in response to the at least one target device control dimension change value being less than the predetermined device control dimension change threshold, maintaining the second device control information to be adjusted unchanged.

According to an embodiment of the present disclosure, the target device control information may be determined as the third device control information to be adjusted in response to detecting that the current environment value is greater than or equal to the predetermined environment threshold and the at least one target device control dimension value is greater than or equal to the predetermined device control dimension threshold corresponding to the at least one target device control dimension value. And adjusting the third device control information to be adjusted in response to the current environmental change value being greater than or equal to the predetermined environmental change threshold and the at least one target device control dimension change value being greater than or equal to the predetermined device control dimension change threshold.

According to the embodiments of the present disclosure, since it is possible to detect whether or not there is at least one of the current environment information satisfying the predetermined environment condition and the target device control information satisfying the predetermined control condition at a predetermined period. And under the condition that the current environment information meets at least one of the preset environment condition and the target equipment control information meets the preset control condition, the target equipment control information is adjusted, so that the timing task triggered based on the condition is realized, various possible conditions can be dealt with, and the robustness and the universality of the system control method are enhanced.

Fig. 4A schematically illustrates an example schematic of a refrigeration system according to an embodiment of the disclosure.

As shown in fig. 4A, a refrigeration system and end device 404 may be included in 400A. The refrigeration system may include a chilled water system 401, a refrigeration unit system 402, and a chilled water system 403. Chilled water system 401 may include a chilled pump 4011. The refrigeration unit system 402 may include a compressor 4021, an evaporator 4022, a condenser 4023, and a throttling element 4024. The cooling water system 403 may include a cooling pump 4031 and a cooling tower 4032.

The heat exchange process of the refrigeration system may include an evaporation heat absorption process of the refrigerant in the evaporator 4022, a condensation heat release process of the refrigerant in the condenser 4023, an absorption process of the chilled water through the terminal equipment 404, and a heat release process of the cooling water through the cooling tower 4032.

In the refrigeration unit system 402, the compressor 4021 may compress low-temperature low-pressure refrigerant vapor into high-temperature high-pressure refrigerant vapor, the high-temperature high-pressure refrigerant vapor enters the condenser 4023, is cooled and released in the condenser 4023, condenses into high-temperature high-pressure refrigerant liquid, is changed into low-temperature low-pressure refrigerant liquid through the throttling element 4024, and finally, the heat of the water return absorbed by the evaporator 4022 is changed into low-temperature low-pressure vapor, and the low-temperature low-pressure vapor enters the compressor 4021 again to start a new refrigeration cycle.

In the cooling water system 403, after absorbing the heat of the condenser 4023, the cooling water increases in temperature and then sprays, and in the spraying process, the cooling water exchanges heat, and then gathers in the water accumulation tray, and then enters the next cycle through the cooling pump 4031. The cooling tower 4032 cools the cooling water having absorbed heat and raised temperature in the condenser 4023, and sends the heat absorbed by the cooling water to the outdoor environment by at least one of convection heat transfer and radiation heat transfer, and the low-temperature cooling water formed after cooling returns to the condenser 4023 to continue the heat absorption cycle.

In the chilled water system 401, low-temperature chilled water produced by the refrigeration unit system 402 is delivered to the terminal equipment 404 corresponding to the refrigeration system by the chilled pump 4011, and chilled water return water after releasing cold energy is returned to the evaporator 4022 of the refrigeration unit system 402.

Fig. 4B schematically illustrates a schematic diagram of determining N levels of candidate device energy consumption impact information corresponding to first candidate operating state information according to the first candidate operating state information in the case where the target system is a refrigeration system according to an embodiment of the present disclosure.

In 400B, N levels of candidate device energy consumption impact information corresponding to the first candidate operating state information may be determined from the first candidate operating state information for the refrigeration system of fig. 4A. N=3.

For level 1, the first candidate operating state information 405 may be subjected to timing feature extraction to obtain candidate operating state feature information 406. The first candidate operating state information 405 may include first candidate device control information and current environment information. The first candidate device control information may include a candidate cooling pump frequency and a candidate cooling tower fan frequency. And obtaining candidate cooling water flow 407 according to the candidate operation state characteristic information 406. The candidate cooling water flow 407 is the 1 st candidate device energy consumption influence information.

For level 2, the first candidate operating state information 405, the candidate operating state feature information 406, and the candidate cooling water flow 407 may be fused to obtain the 1 st fusion information 408. Based on the 1 st fusion information 408, a candidate cooling side inlet/outlet water temperature 409 is obtained. The candidate cooling side inlet water temperature 409 may include a candidate cooling side inlet water temperature and a candidate cooling side outlet water temperature. The candidate cooling side water outlet temperature 409 is the energy consumption influence information of the 2 nd candidate device.

For level 3, the 1 st fusion information 408, the candidate operating state characteristic information 406, and the candidate cooling side water inlet/outlet temperature 409 may be fused to obtain the 2 nd fusion information 410. Candidate cold energy information 411 may be obtained from the 2 nd fusion information 410. The candidate cold energy consumption information 411 is the 3 rd candidate device energy consumption influence information.

In the technical scheme of the disclosure, the related processes of collecting, storing, using, processing, transmitting, providing, disclosing and the like of the personal information of the user accord with the regulations of related laws and regulations, and the public order colloquial is not violated.

In the technical scheme of the disclosure, the authorization or consent of the user is obtained before the personal information of the user is obtained or acquired.

The above is only an exemplary embodiment, but is not limited thereto, and other system control methods known in the art may be included as long as the energy saving effect can be improved.

Fig. 5 schematically illustrates a block diagram of a system control device according to an embodiment of the present disclosure.

As shown in fig. 5, the system control device 500 may include a first determination module 510, a second determination module 520, and a control module 530.

The first determining module 510 is configured to determine, according to at least one first candidate operation state information of the target system, N levels of candidate device energy consumption influence information corresponding to the at least one candidate operation state information.

The second determining module 520 is configured to determine target device control information that satisfies a predetermined power saving condition from the current device control information of the current operation state information and the first candidate device control information of the at least one first candidate operation state information according to the first current selection information corresponding to the current operation state information and the first candidate selection information corresponding to the at least one first candidate operation state information.

And the control module 530 is configured to control the target system to operate according to the target device control information.

According to an embodiment of the present disclosure, the first candidate operation state information includes first candidate device control information and current environment information. The candidate device energy consumption influence information of the current level is associated with the candidate device energy consumption influence information and the input information of the previous level. The input information of the previous level is used for determining candidate equipment energy consumption influence information of the previous level, and the input information of the previous level is determined according to the first candidate running state information. N is an integer greater than 1. The first candidate selection information includes candidate device energy consumption impact information of at least one hierarchy.

According to an embodiment of the present disclosure, the first determination module 510 may include a first determination sub-module and a first obtaining sub-module. The first determining submodule and the first obtaining submodule are used for repeatedly executing the following operations until the N candidate equipment energy consumption influence information is obtained:

the first determining sub-module is used for determining nth fusion information corresponding to at least one first candidate operation state information of the target system according to the candidate operation state characteristic information, the nth-1 fusion information and the nth-1 candidate equipment energy consumption influence information corresponding to the at least one first candidate operation state information. Candidate operating state characteristic information corresponding to the first candidate operating state information is determined based on the first candidate operating state information. N is an integer greater than 1 and less than or equal to N.

The first obtaining sub-module is used for obtaining the n candidate equipment energy consumption influence information corresponding to the at least one first candidate running state information according to the n fusion information corresponding to the at least one first candidate running state information. The n-th candidate device energy consumption influence information characterizes the n-th hierarchical candidate device energy consumption influence information.

According to an embodiment of the present disclosure, the system control device 500 may further include a first obtaining module.

The first obtaining module is used for extracting time sequence characteristics of at least one first candidate operation state information to obtain candidate operation state characteristic information corresponding to the at least one first candidate operation state information.

According to an embodiment of the present disclosure, the at least one first candidate operating state information includes M. The mth first candidate operating state information includes T first candidate operating state sub-information. M is an integer greater than or equal to 1. T is an integer greater than 1. M is an integer greater than or equal to 1 and less than or equal to M.

According to an embodiment of the present disclosure, the first obtaining module may include a second determining sub-module, a third determining sub-module, a fourth determining sub-module, and a fifth determining sub-module.

In response to 1 < t.ltoreq.T,

The second determining sub-module is used for determining forgetting characteristic information of the t first candidate running state sub-information, first updating characteristic information of the t first candidate running state sub-information and candidate state characteristic information of the t first candidate running state sub-information according to the t first candidate running state sub-information and the first hidden characteristic information of the t-1 first candidate running state sub-information.

The third determining sub-module is used for determining the first state characteristic information of the t first candidate operation state sub-information according to the first state characteristic information of the t-1 first candidate operation state sub-information, the forgetting characteristic information of the t first candidate operation state sub-information, the first updated characteristic information of the t first candidate operation state sub-information and the candidate state characteristic information of the t first candidate operation state sub-information.

And the fourth determining sub-module is used for determining the first hidden characteristic information of the t first candidate running state sub-information according to the t first candidate running state sub-information, the first hidden characteristic information of the t-1 first candidate running state sub-information and the first state characteristic information of the t first candidate running state sub-information.

A fifth determining sub-module, configured to determine mth candidate operation state feature information according to the first hidden feature information of the mth first candidate operation state sub-information, where the mth candidate operation state feature information characterizes candidate operation state feature information of the mth first candidate operation state information;

according to an embodiment of the present disclosure, T is an integer greater than or equal to 1 and less than or equal to T. U is an integer greater than or equal to 1 and less than or equal to T.

According to an embodiment of the present disclosure, the at least one first candidate operating state information includes P. The p-th first candidate operating state information includes Q first candidate operating state sub-information. P is an integer greater than or equal to 1. Q is an integer greater than 1. P is an integer greater than or equal to 1 and less than or equal to P.

According to an embodiment of the present disclosure, the first obtaining module may include a sixth determining sub-module, a seventh determining sub-module, an eighth determining sub-module, and a ninth determining sub-module.

In response to 1 < q.ltoreq.Q,

and the sixth determining sub-module is used for determining second updated characteristic information of the first candidate running state sub-information and reset characteristic information of the first candidate running state sub-information according to the second hidden characteristic information of the first candidate running state sub-information and the first candidate running state sub-information.

A seventh determining sub-module, configured to determine second state feature information of the q-th first candidate operating state sub-information according to the q-th first candidate operating state sub-information, the second hidden feature information of the q-1-th first candidate operating state sub-information, and the reset feature information of the q-th first candidate operating state sub-information.

An eighth determining sub-module, configured to determine second hidden feature information of the q-th first candidate running state sub-information according to the second updated feature information of the q-th first candidate running state sub-information, the second hidden feature information of the q-1-th first candidate running state sub-information, and the second state feature information of the q-th first candidate running state sub-information.

And a ninth determining sub-module, configured to determine the p-th candidate operating state feature information according to the second hidden feature information of the V-th first candidate operating state sub-information. The p-th candidate operating state characteristic information characterizes candidate operating state characteristic information of the p-th first candidate operating state information.

According to an embodiment of the present disclosure, Q is an integer greater than 1 or equal to 1 and less than or equal to Q. V is an integer greater than or equal to 1 and less than or equal to Q.

According to an embodiment of the present disclosure, the first obtaining module may include a second obtaining sub-module.

And the second obtaining submodule is used for carrying out R-level processing on the at least one first candidate running state information based on the self-attention strategy to obtain candidate running state characteristic information corresponding to the at least one first candidate running state information. R is an integer greater than or equal to 1.

According to an embodiment of the present disclosure, in the case where R is an integer greater than 1, the second obtaining sub-module may include a first obtaining unit, a second obtaining unit, and a third obtaining unit.

In response to 1 < r.ltoreq.R,

the first obtaining unit is used for obtaining second intermediate candidate running state characteristic information of the r-1 level, which corresponds to the first candidate running state information, according to the first intermediate candidate running state characteristic information of the r-1 level, which corresponds to the first candidate running state information.

The second obtaining unit is configured to obtain first intermediate candidate running state feature information corresponding to the first candidate running state information of the r-th level according to the second intermediate candidate running state feature information corresponding to the first candidate running state information of the r-th level and the first intermediate candidate running state feature information corresponding to the first candidate running state information of the r-1-th level.

And the third obtaining unit is used for obtaining the candidate running state characteristic information corresponding to the first candidate running state information according to the first intermediate candidate running state characteristic information corresponding to the first candidate running state information of the S level.

According to an embodiment of the present disclosure, R is an integer greater than or equal to 1 and less than or equal to R. S is an integer greater than or equal to 1 and less than or equal to R.

According to an embodiment of the present disclosure, the first obtaining module may include a third obtaining sub-module, a fourth obtaining sub-module, and a fifth obtaining sub-module.

And the third obtaining sub-module is used for processing the at least one first candidate running state information based on the sparse self-attention strategy to obtain third intermediate candidate running state characteristic information corresponding to the at least one first candidate running state information.

And a fourth obtaining sub-module, configured to process the third intermediate candidate operating state feature information corresponding to the at least one first candidate operating state information based on the self-focusing distillation policy, to obtain fourth intermediate candidate operating state feature information corresponding to the at least one first candidate operating state information.

And a fifth obtaining sub-module, configured to obtain candidate operation state feature information corresponding to the at least one first candidate operation state information according to the at least one first candidate operation state information and fourth candidate operation state feature information corresponding to the at least one first candidate operation state information.

According to an embodiment of the present disclosure, in the case where J is an integer greater than 1, the first obtaining module may include a sixth obtaining sub-module, a seventh obtaining sub-module, an eighth obtaining sub-module, a ninth obtaining sub-module, and a tenth obtaining sub-module.

In response to 1 < j.ltoreq.J,

and a sixth obtaining sub-module, configured to process the first seasonal feature information corresponding to the first candidate running state information in the j-1 th level based on the autocorrelation policy, to obtain autocorrelation feature information corresponding to the first candidate running state information in the j-1 th level.

And a seventh obtaining sub-module, configured to obtain fifth intermediate candidate running state feature information corresponding to the first candidate running state information of the j-th level according to the autocorrelation feature information corresponding to the first candidate running state information of the j-th level and the first seasonal feature information corresponding to the first candidate running state information of the j-1-th level.

And the eighth obtaining submodule is used for processing fifth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the j-th level based on the first sequence decomposition strategy to obtain second seasonal characteristic information corresponding to the first candidate running state information of the j-th level.

And the ninth obtaining submodule is used for obtaining the first seasonal characteristic information of the j-th level corresponding to the first candidate running state information according to the second seasonal characteristic information of the j-th level corresponding to the first candidate running state information.

A tenth obtaining sub-module, configured to obtain candidate running state feature information corresponding to the first candidate running state information according to first seasonal feature information corresponding to the first candidate running state information of the K-th level;

according to an embodiment of the present disclosure, J is an integer greater than or equal to 1 and less than or equal to J. K is an integer greater than or equal to 1 and less than or equal to J.

According to an embodiment of the present disclosure, the ninth obtaining sub-module may include a fourth obtaining unit, a fifth obtaining unit, and a sixth obtaining unit.

And a fourth obtaining unit, configured to obtain sixth intermediate candidate running state feature information corresponding to the first candidate running state information of the jth level according to the second seasonal feature information corresponding to the first candidate running state information of the jth level.

A fifth obtaining unit, configured to obtain, according to the second seasonal feature information corresponding to the first candidate running state information at the j-th level and the sixth intermediate candidate running state feature information corresponding to the first candidate running state information at the j-th level, seventh intermediate candidate running state feature information corresponding to the first candidate running state information at the j-th level.

And a sixth obtaining unit, configured to process the seventh intermediate candidate running state feature information corresponding to the first candidate running state information in the j-th level based on the second sequence decomposition policy, to obtain first seasonal feature information corresponding to the first candidate running state information in the j-th level.

According to an embodiment of the present disclosure, in the case where W is an integer greater than 1, the first obtaining module may include an eleventh obtaining sub-module, a twelfth obtaining sub-module, a thirteenth obtaining sub-module, a fourteenth obtaining sub-module, and a fifteenth obtaining sub-module.

In response to 1 < w.ltoreq.W,

an eleventh obtaining sub-module, configured to obtain, for the eighth intermediate candidate running state feature information corresponding to the first candidate running state information in the w-1 th level, ninth intermediate candidate running state feature information corresponding to the first candidate running state information in the w-1 th level based on the locally sensitive hash attention policy.

And a twelfth obtaining sub-module, configured to obtain tenth intermediate candidate running state feature information corresponding to the first candidate running state information of the w-th level according to the ninth intermediate candidate running state feature information corresponding to the first candidate running state information of the w-th level and the eighth intermediate candidate running state feature information corresponding to the first candidate running state information of the w-1-th level.

And the thirteenth obtaining sub-module is used for processing tenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level based on the first reversible residual strategy to obtain tenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level.

And a fourteenth obtaining sub-module, configured to obtain, according to the tenth intermediate candidate running state feature information corresponding to the first candidate running state information in the w th level, eighth intermediate candidate running state feature information corresponding to the first candidate running state information in the w th level.

The fifteenth obtaining sub-module is configured to obtain candidate operation state feature information corresponding to the first candidate operation state information according to eighth intermediate candidate operation state feature information corresponding to the first candidate operation state information of the O-th hierarchy.

According to an embodiment of the present disclosure, W is an integer greater than or equal to 1 and less than or equal to W. O is an integer greater than or equal to 1 and less than or equal to W.

According to an embodiment of the present disclosure, the fourteenth obtaining sub-module may include a seventh obtaining unit, an eighth obtaining unit, and a ninth obtaining unit.

A seventh obtaining unit, configured to obtain, according to tenth intermediate candidate operation state feature information corresponding to the first candidate operation state information in the w-th hierarchy, eleventh intermediate candidate operation state feature information corresponding to the first candidate operation state information in the w-th hierarchy.

An eighth obtaining unit, configured to obtain twelfth intermediate candidate operation state feature information corresponding to the first candidate operation state information according to the tenth intermediate candidate operation state feature information corresponding to the first candidate operation state information at the w-th level and the eleventh intermediate candidate operation state feature information corresponding to the first candidate operation state information at the w-th level.

And the ninth obtaining unit is used for processing the twelfth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level based on the second reversible residual strategy to obtain the eighth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level.

According to an embodiment of the present disclosure, in the case where I is an integer greater than 1, the first obtaining unit may include a tenth obtaining unit, an eleventh obtaining unit, a twelfth obtaining unit

In response to 1 < i.ltoreq.I,

and a tenth obtaining unit, configured to process third seasonal feature information corresponding to the first candidate running state information in the i-1 th level based on the frequency domain enhancement policy, so as to obtain frequency domain feature information corresponding to the first candidate running state information in the i-1 th level.

An eleventh obtaining unit, configured to obtain thirteenth intermediate candidate operation state feature information corresponding to the first candidate operation state information of the ith hierarchy according to the frequency domain feature information corresponding to the first candidate operation state information of the ith hierarchy and the third seasonal feature information corresponding to the first candidate operation state information of the ith-1 hierarchy.

A twelfth obtaining unit, configured to process thirteenth intermediate candidate running state feature information corresponding to the first candidate running state information of the ith level based on the first hybrid expert decomposition strategy, to obtain fourth segment feature information corresponding to the first candidate running state information of the ith level.

A thirteenth obtaining unit, configured to obtain third seasonal feature information corresponding to the first candidate running state information of the ith hierarchy according to fourth seasonal feature information corresponding to the first candidate running state information of the ith hierarchy.

A fourteenth obtaining unit, configured to obtain candidate operation state feature information corresponding to the first candidate operation state information according to third seasonal feature information corresponding to the first candidate operation state information of the G-th hierarchy.

According to an embodiment of the present disclosure, I is an integer greater than or equal to 1 and less than or equal to I. G is an integer greater than or equal to 1 and less than or equal to I.

According to an embodiment of the present disclosure, the thirteenth obtaining unit may include a fifteenth obtaining unit, a sixteenth obtaining unit, and a seventeenth obtaining unit.

A fifteenth obtaining unit, configured to obtain fourteenth intermediate candidate operation state feature information corresponding to the first candidate operation state information according to fourth segment feature information corresponding to the first candidate operation state information of the ith hierarchy.

A sixteenth obtaining unit, configured to obtain fifteenth intermediate candidate running state feature information corresponding to the first candidate running state information according to fourth segment feature information corresponding to the first candidate running state information of the ith hierarchy and fourteenth intermediate candidate running state feature information corresponding to the first candidate running state information of the ith hierarchy.

A seventeenth obtaining unit, configured to process fifteenth intermediate candidate running state feature information corresponding to the first candidate running state information in the ith level based on the second hybrid expert decomposition strategy, to obtain third seasonal feature information corresponding to the first candidate running state information in the ith level.

According to an embodiment of the present disclosure, the first determination sub-module may include an eighteenth obtaining unit.

An eighteenth obtaining unit, configured to perform fusion processing on candidate operation state feature information, n-1 th fusion information, and n-1 st candidate device energy consumption influence information corresponding to at least one first candidate operation state information of the target system, to obtain nth fusion information corresponding to the at least one first candidate operation state information.

According to an embodiment of the present disclosure, the first obtaining sub-module may include a nineteenth obtaining unit.

A nineteenth obtaining unit, configured to process, based on the full connection policy, the nth fusion information corresponding to the at least one first candidate operation state information, to obtain nth candidate device energy consumption influence information corresponding to the at least one first candidate operation state information.

According to an embodiment of the present disclosure, the second determination module 520 may include a tenth determination sub-module and an eleventh determination sub-module.

A tenth determination sub-module for determining at least one second candidate operating state information from the at least one first candidate operating state information based on the candidate demand information corresponding to the at least one first candidate operating state information.

An eleventh determination submodule is configured to determine target device control information satisfying a predetermined energy saving condition from current device control information of the current operating state information and first candidate device control information of the at least one second candidate operating state information, based on the current device energy consumption information corresponding to the current operating state information and the candidate device energy consumption information corresponding to the at least one second candidate operating state information.

According to an embodiment of the present disclosure, the first current selection information further includes current device operation mode information. The first candidate selection information also includes candidate device operating mode information.

According to an embodiment of the present disclosure, the eleventh determination sub-module may include a first determination unit and a second determination unit.

The first determining unit is used for determining at least one third candidate operation state information from the at least one second candidate operation state information according to the current device operation mode information corresponding to the current operation state information and the candidate device operation mode information corresponding to the at least one second candidate operation state information.

And a second determining unit configured to determine target device control information satisfying a predetermined energy saving condition from the current device control information of the current operation state information and the first candidate device control information of the at least one third candidate operation state information, based on the current device energy consumption information corresponding to the current operation state information and the candidate device energy consumption information corresponding to the at least one third candidate operation state information.

According to an embodiment of the present disclosure, the system control device 500 may further include a third determination module and a fourth determination module.

And a third determining module for determining at least one second candidate device control information from the plurality of historical device control information of the target system.

And a fourth determining module, configured to determine first candidate device control information of the at least one first candidate operation state information according to the at least one second candidate device control information.

According to an embodiment of the present disclosure, the third determination module may include a twelfth determination sub-module.

A twelfth determining sub-module for determining at least one second candidate device control information from the plurality of historical device control information according to the second current selection information corresponding to the current device control information and the second candidate selection information corresponding to the plurality of historical device control information of the target system.

According to an embodiment of the present disclosure, the second current selection information includes current load information and current environment information. The second candidate selection information includes historical load information and historical environment information.

According to an embodiment of the present disclosure, the twelfth determination sub-module may include a third determination unit and a fourth determination unit.

And a third determining unit configured to determine at least one third candidate device control information from the plurality of historical device control information based on the current load information corresponding to the current device control information and the historical load information corresponding to the plurality of historical device control information of the target system.

And a fourth determining unit configured to determine at least one second candidate device control information from the at least one third candidate device control information based on the current environment information corresponding to the current device control information and the historical environment information corresponding to the at least one third candidate device control information.

According to an embodiment of the present disclosure, the fourth determination unit may include a first determination subunit and a second determination subunit.

And the first determining subunit is used for determining the similarity between the current environment information corresponding to the current device control information and the historical environment information corresponding to the at least one third candidate device control information to obtain at least one similarity.

And a second determining subunit configured to determine at least one second candidate device control information from the at least one third candidate device control information according to the at least one similarity.

According to an embodiment of the present disclosure, the first candidate run state information includes first candidate run state dimension information of at least one dimension.

According to an embodiment of the present disclosure, the system control device 500 may further include a fifth determination module, a sixth determination module, and a seventh determination module.

And the fifth determining module is used for obtaining the historical running state characteristic information corresponding to the plurality of candidate dimensions according to the historical running state information corresponding to the plurality of candidate dimensions.

And the sixth determining module is used for determining importance degrees corresponding to the plurality of candidate dimensions according to the historical operation state characteristic information corresponding to the plurality of candidate dimensions.

And a seventh determining module, configured to determine at least one dimension from the plurality of candidate dimensions according to importance levels corresponding to the plurality of candidate dimensions.

According to an embodiment of the disclosure, the fifth determination module may include a sixteenth acquisition sub-module

And a sixteenth obtaining sub-module, configured to process the historical running state information corresponding to the multiple candidate dimensions by using the characterization model, so as to obtain historical running state feature information corresponding to the multiple candidate dimensions.

According to an embodiment of the present disclosure, the characterization model is derived by training a self-supervision model using loss function values. The loss function value is based on the loss function. And determining according to the sample operation state characteristic information of the positive sample and the sample operation state characteristic information of a plurality of negative samples corresponding to the positive sample.

According to an embodiment of the present disclosure, the plurality of negative samples corresponding to the positive sample are determined from the plurality of candidate negative samples according to sample operation state characteristic information of the positive sample and sample operation state characteristic information of the plurality of candidate negative samples corresponding to the positive sample.

According to an embodiment of the present disclosure, the sample operation state characteristic information of the positive sample is obtained by processing the positive sample using a self-supervision model.

According to an embodiment of the present disclosure, the sample operating state characteristic information of the negative sample is obtained by processing the negative sample using a self-supervision model.

The system control device 500 may further include an adjustment module according to an embodiment of the present disclosure.

And an adjustment module for adjusting the target device control information in response to detecting that the current environmental information satisfies at least one of a predetermined environmental condition and the target device control information satisfies a predetermined control condition.

According to an embodiment of the present disclosure, the target system includes a refrigeration system.

According to an embodiment of the present disclosure, the first candidate device control information includes a candidate cooling pump frequency and a candidate cooling tower fan frequency.

According to embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

According to an embodiment of the present disclosure, an electronic device includes: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executable by the at least one processor to enable the at least one processor to perform the method as described above.

According to an embodiment of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the method as described above.

According to an embodiment of the present disclosure, a computer program product comprising a computer program which, when executed by a processor, implements a method as described above.

Fig. 6 schematically illustrates a block diagram of an electronic device adapted to implement a system control method according to an embodiment of the disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein.

As shown in fig. 6, the electronic device 600 includes a computing unit 601 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM) 602 or a computer program loaded from a storage unit 608 into a Random Access Memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 600 can also be stored. The computing unit 601, ROM 602, and RAM 603 are connected to each other by a bus 604. An input/output (I/O) interface 605 is also connected to bus 604.

A number of components in the electronic device 600 are connected to the I/O interface 605, including: an input unit 606 such as a keyboard, mouse, etc.; an output unit 607 such as various types of displays, speakers, and the like; a storage unit 608, such as a magnetic disk, optical disk, or the like; and a communication unit 609 such as a network card, modem, wireless communication transceiver, etc. The communication unit 609 allows the electronic device 600 to exchange information/data with other devices through a computer network, such as the internet, and/or various telecommunication networks.

The computing unit 601 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 601 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 601 performs the respective methods and processes described above, for example, a system control method. For example, in some embodiments, the system control method may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as storage unit 608. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device 600 via the ROM 602 and/or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the system control method described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the system control method by any other suitable means (e.g., by means of firmware).

Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), complex Programmable Logic Devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.

Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program code may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus such that the program code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and pointing device (e.g., a mouse or trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), and the internet.

The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server may be a cloud server, a server of a distributed system, or a server incorporating a blockchain.

It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel or sequentially or in a different order, provided that the desired results of the technical solutions of the present disclosure are achieved, and are not limited herein.

The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.

Claims

1. A system control method, comprising:

determining N levels of candidate device energy consumption influence information corresponding to at least one first candidate operation state information of a target system, wherein the target system is a system for generating physical quantities by consuming energy, the target system comprises at least one target device, the first candidate operation state information comprises first candidate device control information and current environment information, the candidate device energy consumption influence information of the current level is associated with candidate device energy consumption influence information of a last level and input information, the input information of the last level is used for determining the candidate device energy consumption influence information of the last level, the input information of the last level is determined according to the first candidate operation state information, and N is an integer greater than 1;

Determining target device control information meeting a predetermined energy saving condition from current device control information of current operation state information and first candidate device control information of at least one first candidate operation state information according to first current selection information corresponding to the current operation state information and first candidate selection information corresponding to the at least one first candidate operation state information, wherein the first current selection information comprises current device energy consumption information and current device operation mode information, the first candidate selection information comprises at least one hierarchy of candidate device energy consumption influence information and candidate device operation mode information, and the candidate device energy consumption influence information comprises candidate device energy consumption information and candidate demand amount information; and

and controlling the target system to run according to the target equipment control information.

2. The method of claim 1, wherein the determining N levels of candidate device energy consumption impact information corresponding to at least one first candidate operating state information of the target system according to the at least one first candidate operating state information comprises repeating the following operations until an nth candidate device energy consumption impact information is obtained:

Determining nth fusion information corresponding to at least one first candidate operating state information of the target system according to the candidate operating state characteristic information, the nth-1 fusion information and the nth-1 candidate equipment energy consumption influence information corresponding to the at least one first candidate operating state information, wherein the candidate operating state characteristic information corresponding to the first candidate operating state information is determined according to the first candidate operating state information, and N is an integer greater than 1 and less than or equal to N; and

according to the nth fusion information corresponding to the at least one first candidate running state information, nth candidate equipment energy consumption influence information corresponding to the at least one first candidate running state information is obtained, wherein the nth candidate equipment energy consumption influence information represents the candidate equipment energy consumption influence information of the nth hierarchy.

3. The method of claim 2, further comprising:

and extracting time sequence characteristics of the at least one first candidate running state information to obtain candidate running state characteristic information corresponding to the at least one first candidate running state information.

4. The method of claim 3, wherein the at least one first candidate operating state information comprises M, the mth first candidate operating state information comprises T first candidate operating state sub-information, M is an integer greater than or equal to 1, T is an integer greater than or equal to 1, and M is an integer greater than or equal to 1 and less than or equal to M;

The step of extracting the time sequence feature of the at least one first candidate operation state information to obtain candidate operation state feature information corresponding to the at least one first candidate operation state information includes:

in response to 1<t being less than or equal to T,

determining forgetting characteristic information of the t-th first candidate running state sub-information, first updating characteristic information of the t-th first candidate running state sub-information and candidate state characteristic information of the t-th first candidate running state sub-information according to the t-th first candidate running state sub-information and the first hidden characteristic information of the t-1 th first candidate running state sub-information;

determining first state characteristic information of the t-th first candidate operation state sub-information according to the first state characteristic information of the t-1-th first candidate operation state sub-information, the forgetting characteristic information of the t-th first candidate operation state sub-information, the first updated characteristic information of the t-th first candidate operation state sub-information and the candidate state characteristic information of the t-th first candidate operation state sub-information;

determining first hidden characteristic information of the t-th first candidate operation state sub-information according to the t-th first candidate operation state sub-information, the t-1-th first hidden characteristic information of the first candidate operation state sub-information and the first state characteristic information of the t-th first candidate operation state sub-information; and

Determining the mth candidate running state characteristic information according to the first hidden characteristic information of the first candidate running state sub-information, wherein the mth candidate running state characteristic information represents the candidate running state characteristic information of the mth candidate running state information;

wherein T is an integer greater than or equal to 1 and less than or equal to T, and U is an integer greater than or equal to 1 and less than or equal to T.

5. The method of claim 3, wherein the at least one first candidate operating state information comprises P, the P-th first candidate operating state information comprises Q first candidate operating state sub-information, P is an integer greater than or equal to 1, Q is an integer greater than 1, and P is an integer greater than or equal to 1 and less than or equal to P;

in response to 1 < q.ltoreq.Q,

determining second updated characteristic information of the first candidate running state sub-information and reset characteristic information of the first candidate running state sub-information according to the second hidden characteristic information of the first candidate running state sub-information and the q-1 th first candidate running state sub-information;

Determining second state characteristic information of the first candidate running state sub-information according to the q first candidate running state sub-information, the second hidden characteristic information of the q-1 first candidate running state sub-information and the reset characteristic information of the q first candidate running state sub-information;

determining second hidden characteristic information of the first candidate running state sub-information according to the second updated characteristic information of the first candidate running state sub-information, the second hidden characteristic information of the q-1 first candidate running state sub-information and the second state characteristic information of the first candidate running state sub-information; and

according to the second hidden characteristic information of the V-th first candidate running state sub-information, determining p-th candidate running state characteristic information, wherein the p-th candidate running state characteristic information represents the p-th candidate running state characteristic information of the first candidate running state information;

wherein Q is an integer greater than or equal to 1 and less than or equal to Q, and V is an integer greater than or equal to 1 and less than or equal to Q.

6. The method of claim 3, wherein the performing timing feature extraction on the at least one first candidate operating state information to obtain candidate operating state feature information corresponding to the at least one first candidate operating state information includes:

And carrying out R layers of processing on the at least one first candidate running state information based on the self-attention strategy to obtain candidate running state characteristic information corresponding to the at least one first candidate running state information, wherein R is an integer greater than or equal to 1.

7. The method according to claim 6, wherein, in the case where R is an integer greater than 1, the performing R-level processing on the at least one first candidate operation state information based on the self-attention policy, to obtain candidate operation state feature information corresponding to the at least one first candidate operation state information, includes:

in response to 1 < r.ltoreq.R,

obtaining second intermediate candidate running state characteristic information of the r-1 level corresponding to the first candidate running state information according to the first intermediate candidate running state characteristic information of the r-1 level corresponding to the first candidate running state information;

obtaining first intermediate candidate running state characteristic information of the r-1 level corresponding to the first candidate running state information according to the second intermediate candidate running state characteristic information of the r-1 level corresponding to the first candidate running state information and the first intermediate candidate running state characteristic information of the r-1 level corresponding to the first candidate running state information; and

Obtaining candidate running state characteristic information corresponding to the first candidate running state information according to first intermediate candidate running state characteristic information corresponding to the first candidate running state information of an S-level;

wherein R is an integer greater than or equal to 1 and less than or equal to R, and S is an integer greater than or equal to 1 and less than or equal to R.

8. The method of claim 3, wherein the performing timing feature extraction on the at least one first candidate operating state information to obtain candidate operating state feature information corresponding to the at least one first candidate operating state information includes:

processing the at least one first candidate running state information based on a sparse self-attention strategy to obtain third intermediate candidate running state characteristic information corresponding to the at least one first candidate running state information;

processing third intermediate candidate operation state characteristic information corresponding to the at least one first candidate operation state information based on a self-focusing distillation strategy to obtain fourth intermediate candidate operation state characteristic information corresponding to the at least one first candidate operation state information; and

And obtaining candidate running state characteristic information corresponding to the at least one first candidate running state information according to the at least one first candidate running state information and fourth intermediate candidate running state characteristic information corresponding to the at least one first candidate running state information.

9. The method according to claim 3, wherein, in the case where J is an integer greater than 1, the performing timing feature extraction on the at least one first candidate operation state information to obtain candidate operation state feature information corresponding to the at least one first candidate operation state information includes:

in response to 1 < j.ltoreq.J,

processing first seasonal feature information corresponding to the first candidate running state information of the j-1 th level based on an autocorrelation strategy to obtain autocorrelation feature information corresponding to the first candidate running state information of the j-1 th level;

obtaining fifth intermediate candidate running state characteristic information of the j-th level, which corresponds to the first candidate running state information, according to the autocorrelation characteristic information of the j-th level, which corresponds to the first candidate running state information, and the first seasonal characteristic information of the j-1-th level, which corresponds to the first candidate running state information;

Processing fifth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the j-th level based on a first sequence decomposition strategy to obtain second seasonal characteristic information corresponding to the first candidate running state information of the j-th level;

obtaining first seasonal characteristic information of the jth level corresponding to the first candidate running state information according to second seasonal characteristic information of the jth level corresponding to the first candidate running state information; and

obtaining candidate running state characteristic information corresponding to the first candidate running state information according to first seasonal characteristic information corresponding to the first candidate running state information of a K-th level;

wherein J is an integer greater than or equal to 1 and less than or equal to J, and K is an integer greater than or equal to 1 and less than or equal to J.

10. The method according to claim 9, wherein the obtaining the first seasonal feature information of the j-th level corresponding to the first candidate operation state information according to the second seasonal feature information of the j-th level corresponding to the first candidate operation state information includes:

obtaining sixth intermediate candidate running state characteristic information of the jth level, which corresponds to the first candidate running state information, according to the second seasonal characteristic information of the jth level, which corresponds to the first candidate running state information;

Obtaining seventh intermediate candidate running state characteristic information of the jth level, which corresponds to the first candidate running state information, according to the second seasonal characteristic information of the jth level, which corresponds to the first candidate running state information, and the sixth intermediate candidate running state characteristic information of the jth level, which corresponds to the first candidate running state information; and

and processing seventh intermediate candidate running state characteristic information corresponding to the first candidate running state information of the jth level based on a second sequence decomposition strategy to obtain first seasonal characteristic information corresponding to the first candidate running state information of the jth level.

11. The method according to claim 3, wherein, in the case where W is an integer greater than 1, the performing timing feature extraction on the at least one first candidate operation state information to obtain candidate operation state feature information corresponding to the at least one first candidate operation state information includes:

in response to 1 < w.ltoreq.W,

obtaining ninth intermediate candidate running state characteristic information of the w-1 level corresponding to the first candidate running state information based on the local sensitive hash attention strategy for the eighth intermediate candidate running state characteristic information of the w-1 level corresponding to the first candidate running state information;

Obtaining tenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level according to the ninth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level and the eighth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-1-th level;

processing tenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level based on a first reversible residual strategy to obtain tenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level;

according to the tenth intermediate candidate running state characteristic information of the w-th level, which corresponds to the first candidate running state information, eighth intermediate candidate running state characteristic information of the w-th level, which corresponds to the first candidate running state information, is obtained; and

obtaining candidate running state characteristic information corresponding to the first candidate running state information according to eighth intermediate running state characteristic information of an O level, which corresponds to the first candidate running state information;

Wherein W is an integer greater than or equal to 1 and less than or equal to W, and O is an integer greater than or equal to 1 and less than or equal to W.

12. The method of claim 11, wherein the obtaining, according to the tenth intermediate candidate operating state characteristic information corresponding to the first candidate operating state information of the w-th hierarchy, eighth intermediate candidate operating state characteristic information corresponding to the first candidate operating state information of the w-th hierarchy includes:

according to tenth intermediate candidate running state characteristic information of the w-th level, which corresponds to the first candidate running state information, eleventh intermediate candidate running state characteristic information of the w-th level, which corresponds to the first candidate running state information, is obtained;

obtaining twelfth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level according to the tenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level and the eleventh intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level; and

and processing twelfth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level based on a second reversible residual strategy to obtain eighth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the w-th level.

13. The method according to claim 3, wherein, in the case where I is an integer greater than 1, the performing timing feature extraction on the at least one first candidate operation state information to obtain candidate operation state feature information corresponding to the at least one first candidate operation state information includes:

in response to 1 < i.ltoreq.I,

processing third seasonal feature information corresponding to the first candidate running state information of the ith hierarchy based on a frequency domain enhancement strategy to obtain frequency domain feature information corresponding to the first candidate running state information of the ith hierarchy;

obtaining thirteenth intermediate candidate operation state characteristic information of the ith hierarchy corresponding to the first candidate operation state information according to the frequency domain characteristic information of the ith hierarchy corresponding to the first candidate operation state information and the third seasonal characteristic information of the ith-1 hierarchy corresponding to the first candidate operation state information;

processing thirteenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the ith hierarchy based on a first mixed expert decomposition strategy to obtain fourth section characteristic information corresponding to the first candidate running state information of the ith hierarchy;

Obtaining third seasonal characteristic information of the ith level corresponding to the first candidate running state information according to the fourth seasonal characteristic information of the ith level corresponding to the first candidate running state information; and

obtaining candidate running state characteristic information corresponding to the first candidate running state information according to third seasonal characteristic information of a G level, which corresponds to the first candidate running state information;

wherein I is an integer greater than or equal to 1 and less than or equal to I, and G is an integer greater than or equal to 1 and less than or equal to I.

14. The method of claim 13, wherein the obtaining third seasonal feature information of the ith hierarchy corresponding to the first candidate operating state information according to fourth seasonal feature information of the ith hierarchy corresponding to the first candidate operating state information includes:

obtaining fourteenth intermediate candidate running state characteristic information of the ith hierarchy corresponding to the first candidate running state information according to fourth season characteristic information of the ith hierarchy corresponding to the first candidate running state information;

obtaining fifteenth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the ith hierarchy according to fourth segment characteristic information corresponding to the first candidate operation state information of the ith hierarchy and fourteenth intermediate candidate operation state characteristic information corresponding to the first candidate operation state information of the ith hierarchy; and

And processing fifteenth intermediate candidate running state characteristic information corresponding to the first candidate running state information of the ith level based on a second mixed expert decomposition strategy to obtain third seasonal characteristic information corresponding to the first candidate running state information of the ith level.

15. The method of any of claims 2-14, wherein the determining the nth fusion information corresponding to the at least one first candidate operating state information according to candidate operating state characteristic information, the nth-1 fusion information, and the nth-1 candidate device energy consumption influence information corresponding to the at least one first candidate operating state information of the target system comprises:

16. The method according to any one of claims 2-14, wherein the obtaining, according to the nth fusion information corresponding to the at least one first candidate operation state information, nth candidate device energy consumption influence information corresponding to the at least one first candidate operation state information includes:

And processing the nth fusion information corresponding to the at least one first candidate running state information based on the full connection strategy to obtain nth candidate equipment energy consumption influence information corresponding to the at least one first candidate running state information.

17. The method of any of claims 1-14, wherein the determining target device control information that satisfies a predetermined power saving condition from current device control information of the current operating state information and first candidate device control information of the at least one first candidate operating state information based on first current selection information corresponding to the current operating state information and first candidate selection information corresponding to the at least one first candidate operating state information comprises:

determining at least one second candidate operation state information from the at least one first candidate operation state information according to the candidate demand amount information corresponding to the at least one first candidate operation state information; and

determining target device control information meeting the predetermined energy saving condition from the current device control information of the current operating state information and the first candidate device control information of the at least one second candidate operating state information according to the current device energy consumption information corresponding to the current operating state information and the candidate device energy consumption information corresponding to the at least one second candidate operating state information.

18. The method of claim 17, wherein the determining target device control information that satisfies the predetermined power saving condition from current device control information of the current operating state information and first candidate device control information of the at least one second candidate operating state information based on the current device power consumption information corresponding to the current operating state information and candidate device power consumption information corresponding to the at least one second candidate operating state information comprises:

determining at least one third candidate operation state information from the at least one second candidate operation state information according to the current device operation mode information corresponding to the current operation state information and the candidate device operation mode information corresponding to the at least one second candidate operation state information; and

determining target device control information meeting the predetermined energy saving condition from the current device control information of the current operating state information and the first candidate device control information of the at least one third candidate operating state information according to the current device energy consumption information corresponding to the current operating state information and the candidate device energy consumption information corresponding to the at least one third candidate operating state information.

19. The method of any one of claims 1-14, further comprising:

determining at least one second candidate device control information from a plurality of historical device control information of the target system; and

and determining first candidate device control information of the at least one first candidate operation state information according to the at least one second candidate device control information.

20. The method of claim 19, wherein the determining at least one second candidate device control information from a plurality of historical device control information for the target system comprises:

and determining at least one piece of second candidate device control information from the plurality of pieces of historical device control information according to second current selection information corresponding to the current device control information and second candidate selection information corresponding to the plurality of pieces of historical device control information of the target system.

21. The method of claim 20, wherein the second current selection information includes current load information and the current context information, and the second candidate selection information includes historical load information and historical context information;

wherein the determining the at least one second candidate device control information from the plurality of historical device control information according to the second current selection information corresponding to the current device control information and the second candidate selection information corresponding to the plurality of historical device control information of the target system includes:

Determining at least one third candidate device control information from the plurality of historical device control information according to current load information corresponding to the current device control information and historical load information corresponding to the plurality of historical device control information of the target system; and

determining the at least one second candidate device control information from the at least one third candidate device control information according to the current environment information corresponding to the current device control information and the historical environment information corresponding to the at least one third candidate device control information.

22. The method of claim 21, wherein the determining the at least one second candidate device control information from the at least one third candidate device control information based on current environmental information corresponding to the current device control information and historical environmental information corresponding to the at least one third candidate device control information comprises:

determining the similarity between the current environmental information corresponding to the current equipment control information and the historical environmental information corresponding to the at least one third candidate equipment control information to obtain at least one similarity; and

And determining the at least one second candidate device control information from the at least one third candidate device control information according to the at least one similarity.

23. The method of any of claims 1-14, wherein the first candidate operating state information includes first candidate operating state dimension information for at least one dimension;

the method further comprises the steps of:

obtaining historical running state characteristic information corresponding to a plurality of candidate dimensions according to the historical running state information corresponding to the plurality of candidate dimensions;

determining importance degrees corresponding to the plurality of candidate dimensions according to the historical operation state characteristic information corresponding to the plurality of candidate dimensions; and

the at least one dimension is determined from the plurality of candidate dimensions according to importance levels corresponding to the plurality of candidate dimensions.

24. The method of claim 23, wherein the deriving historical operating state characteristic information corresponding to the plurality of candidate dimensions from the historical operating state information corresponding to the plurality of candidate dimensions comprises:

processing the historical operation state information corresponding to the plurality of candidate dimensions by using a characterization model to obtain historical operation state characteristic information corresponding to the plurality of candidate dimensions;

The characterization model is obtained by training a self-supervision model by using a loss function value, and the loss function value is determined according to sample running state characteristic information of a positive sample and sample running state characteristic information of a plurality of negative samples corresponding to the positive sample based on the loss function.

25. The method of claim 24, wherein the plurality of negative samples corresponding to the positive sample are determined from the plurality of candidate negative samples based on sample run state characteristic information of the positive sample and sample run state characteristic information of the plurality of candidate negative samples corresponding to the positive sample;

the sample running state characteristic information of the positive sample is obtained by processing the positive sample by using the self-supervision model;

the sample running state characteristic information of the negative sample is obtained by processing the negative sample by using the self-supervision model.

26. The method of any one of claims 1-14, further comprising:

and adjusting the target device control information in response to detecting that the current environmental information meets at least one of a predetermined environmental condition and the target device control information meets a predetermined control condition.

27. The method of any of claims 1-14, wherein the target system comprises a refrigeration system.

28. The method of claim 27, wherein the first candidate plant control information includes a candidate cooling pump frequency and a candidate cooling tower fan frequency.

29. A system control device comprising:

a first determining module, configured to determine N levels of candidate device energy consumption influence information corresponding to at least one candidate operation state information according to at least one first candidate operation state information of a target system, where the target system is a system that generates a physical quantity by consuming energy, the target system includes at least one target device, the first candidate operation state information includes first candidate device control information and current environment information, the candidate device energy consumption influence information of the current level is associated with candidate device energy consumption influence information of a previous level and input information, the input information of the previous level is used to determine candidate device energy consumption influence information of the previous level, the input information of the previous level is determined according to the first candidate operation state information, and N is an integer greater than 1;

A second determining module, configured to determine target device control information that meets a predetermined energy saving condition from current device control information of current operation state information and first candidate device control information of at least one first candidate operation state information according to first current selection information corresponding to the current operation state information and first candidate selection information corresponding to the at least one first candidate operation state information, where the first current selection information includes current device energy consumption information and current device operation mode information, and the first candidate selection information includes at least one hierarchy of the candidate device energy consumption influence information and candidate device operation mode information, and the candidate device energy consumption influence information includes candidate device energy consumption information and candidate demand amount information; and

and the control module is used for controlling the operation of the target system according to the control information of the target equipment.

30. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 28.

31. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-28.

32. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 28.