CN107770783B

CN107770783B - Base station capacity expansion transformation scheme design method and related equipment

Info

Publication number: CN107770783B
Application number: CN201710924326.7A
Authority: CN
Inventors: 刘颖; 乔健; 黄飞
Original assignee: Huawei Technologies Co Ltd
Current assignee: Huawei Technologies Co Ltd
Priority date: 2017-09-30
Filing date: 2017-09-30
Publication date: 2020-10-09
Anticipated expiration: 2037-09-30
Also published as: CN107770783A

Abstract

The application discloses a base station capacity expansion transformation scheme design method and related equipment, which are used for improving the speed, efficiency and accuracy of base station scheme design to be designed. The method in the embodiment of the application comprises the following steps: acquiring information of a current network base station, information of a base station to be designed and a historical scheme set; carrying out data processing on the information of the current network base station and the information of the base station to be designed to obtain the characteristics to be used; carrying out algorithm training learning on the historical scheme set to obtain a scheme evolution path model and a device connection knowledge map; selecting a candidate scheme subset from the historical scheme set through a reverse subtraction algorithm according to the features to be used; performing recommendation degree sorting processing on all candidate schemes in the candidate scheme subset according to the scheme evolution path model to obtain a recommendation result; and when the incomplete candidate scheme exists in the recommendation result, performing self-perfection processing on the incomplete candidate scheme according to the device connection knowledge graph to obtain a base station scheme set to be designed.

Description

Base station capacity expansion transformation scheme design method and related equipment

Technical Field

The present application relates to the field of communications technologies, and in particular, to a method for designing a base station capacity expansion modification scheme and a related device.

Background

With the gradual evolution of the global network towards 4G and 5G, the purely new scenes are less and less, the capacity expansion and the reconstruction become mainstream, and the investment and the difficulty of the base station design are greater and greater.

The deployment scheme under the base station capacity expansion and transformation scene needs to be designed based on the existing network stock, the design process is completely finished manually, and the difference lies in the graph mode. The method I is to manufacture PPT (PowerPoint) for pure manual work, and specifically comprises the following steps: based on customer requirements, a Bill Of Quotation (BOQ) and current network inventory configuration data, PPT is manually made according to expert experience. The second mode is that the tool is automatically drawn, and specifically: importing existing network stock configuration data, and finishing the restoration of an existing network base station deployment scheme by using a tool; based on the current network scheme and the configuration information of products before sale, and depending on expert experience, a target network scheme meeting Radio Frequency (RF) planning is designed manually; and filling and importing a main device wiring rule table and a site information table (BOQ and RF planning information) based on the design result, reading parameters by the tool, and automatically generating a main device planning diagram based on the built-in hardware rule.

The first and second methods have the following disadvantages: 1. the capacity expansion transformation is carried out in a forward mode, in the scheme design, the types of the devices are more, the quantity selection is not fixed, the slot position selection of the devices and the port connection mode selection between the devices are more, the combination explosion occurs, namely the number of the permutation and combination is too large, the quantity of the candidate schemes which can be generated is huge, and the final scheme is difficult to judge and select by manpower from many combination designs; 2. the design of the deployment scheme of the capacity expansion reconstruction scene needs to refer to the existing network scheme, the design process is complex, and the experience requirement on technicians is higher than that of a newly-built scene; 3. the design link is completely completed by depending on expert experience, so that the time is consumed and the error rate is high; 4. the current scheme does not realize cloud storage, and the optimal scheme cannot be shared on the global scale. In the above four points, the speed and efficiency of designing the capacity expansion and transformation scheme of the base station are low.

Disclosure of Invention

The application provides a base station capacity expansion transformation scheme design method and related equipment, which are used for improving the speed, efficiency and accuracy of base station scheme design to be designed.

A first aspect of the present application provides a method for designing a capacity expansion and reconstruction scheme of a base station, including:

acquiring information of a current network base station, information of a base station to be designed and a historical scheme set;

performing data processing on the information of the current network base station and the information of the base station to be designed to obtain the characteristics to be used;

carrying out algorithm training learning on the historical scheme set to obtain a scheme evolution path model and a device connection knowledge map;

according to the features to be used, selecting a candidate scheme subset from the historical scheme set through a reverse subtraction algorithm;

performing recommendation degree sorting processing on all candidate schemes in the candidate scheme subset according to the scheme evolution path model to obtain a recommendation result;

and when an incomplete candidate scheme exists in the recommendation result, performing self-perfection processing on the incomplete candidate scheme according to the device connection knowledge graph to obtain a base station scheme set to be designed.

When a base station needs to be subjected to capacity expansion modification, a base station design delivery person inputs current network base station information and base station information to be designed into a base station capacity expansion modification scheme design device, the base station capacity expansion modification scheme design device can acquire the current network base station information and the base station information to be designed input by the base station design delivery person, acquire a historical scheme set from a network or a database of the device through the input of the base station design delivery person, perform algorithm training and learning on the historical scheme set to obtain a scheme evolution path model and a device connection knowledge map, perform feature matching of features to be used on the historical scheme set through a reverse subtraction algorithm according to the features to be used so as to select a candidate scheme subset, and can deduce and obtain an evolution path of a capacity expansion modification scheme of the base station to be designed on the basis of the scheme evolution path model, and when the incomplete candidate scheme exists in the recommendation result, the connection rule among the devices can be known according to the device connection knowledge map, the incomplete candidate scheme is subjected to self-perfection according to the connection rule among the devices, so that the incomplete candidate scheme can be completely matched with the base station to be designed, and the incomplete candidate scheme after the self-perfection and the originally perfect candidate scheme are combined into a set to obtain a base station scheme set to be designed. By using a reverse subtraction algorithm to obtain the candidate scheme subset, the problem of combined explosion caused by excessive scenes in forward evolution logic can be avoided, and the number of candidate schemes is obviously reduced; performing recommendation degree sorting processing on all candidate schemes in the candidate scheme subset according to the scheme evolution path model to obtain a recommendation result, so that the candidate schemes can be obtained more accurately and quickly; and performing self-perfection processing on the imperfect candidate scheme according to the device connection knowledge graph, so that the candidate scheme which better meets the requirements of customers can be provided. Therefore, the speed, the efficiency and the accuracy of the scheme design of the base station to be designed are improved.

With reference to the first aspect of the present application, in a first implementation manner of the first aspect of the present application, the performing algorithm training and learning on the historical solution set to obtain an evolution knowledge graph and a device connection knowledge graph includes:

carrying out evolution path training learning on the historical scheme set through a social graph mining algorithm to obtain a scheme evolution path model;

and performing connection knowledge graph training learning on the historical scheme set through a frequent item mining algorithm to obtain a device connection knowledge graph.

The training learning algorithm of the scheme evolution path model may be obtained through a social graph mining algorithm, and the scheme evolution path model is mainly used for deducing a connection relationship between devices of each capacity expansion transformation scheme in the historical scheme set. The key points of the training and learning algorithm of the device connection knowledge graph are to learn the basic connection rules and deep connection relations among the devices and the network structure and parameters of the probability graph model. And obtaining a frequent association sequence and a probability graph model through a frequent item mining algorithm, and obtaining the device connection knowledge graph on the basis of the frequent association sequence and the probability graph model.

With reference to the first embodiment of the first aspect of the present application, in a second embodiment of the first aspect of the present application, the selecting, according to the feature to be used, a candidate scheme subset from the historical scheme set by using a reverse subtraction algorithm includes:

obtaining an RF characteristic, a new BOQ characteristic, an existing network BOQ characteristic and a hardware replacement condition characteristic according to the characteristics to be used;

performing feature matching on all candidate schemes in the historical scheme set according to the RF features and the new BOQ features to obtain a first subset;

judging whether the hardware is replaceable according to the hardware replacement condition characteristics;

if the hardware replacement condition features are not hardware replaceable, performing feature matching on all candidate schemes in the first subset according to the current network BOQ features to obtain a second subset;

calculating the feature matching degree of each candidate scheme in the second subset, and sequencing the schemes in the second subset according to the feature matching degree to obtain a candidate scheme subset;

if the hardware replacement condition characteristic is hardware replacement, deleting all candidate schemes in the first subset according to the new BOQ characteristic to obtain a third subset, wherein all candidate schemes in the third subset do not contain the new BOQ characteristic;

performing feature matching on all candidate schemes in the third subset according to the current network BOQ features to obtain a fourth subset;

and calculating the feature matching degree of each candidate scheme in the fourth subset, and sequencing the candidate schemes in the fourth subset according to the feature matching degree to obtain a candidate scheme subset.

According to the characteristics to be used, obtaining RF characteristics, new BOQ characteristics, current network BOQ characteristics and hardware replacement condition characteristics, wherein the RF characteristics and the new BOQ are important data for carrying out capacity expansion and reconstruction on the base station to be designed, the hardware replacement condition characteristics are preset by base station design delivery personnel and are divided into hardware replacement and hardware non-replacement, a first subset is obtained from matching the RF signature with the new BOQ signature against all of the plans in the historical plan library, judging whether the hardware is replaceable according to the hardware replacement condition characteristic, if the hardware replacement condition characteristic is not replaceable, according to the current network BOQ characteristic, performing feature matching on all candidate schemes in the first subset to obtain a second subset, calculating the feature matching degree of each candidate scheme in the second subset, sorting the schemes in the second subset according to the feature matching degree to obtain a candidate scheme subset; if the hardware replacement condition features are hardware replacement, deleting and selecting all candidate schemes in the first subset according to the new BOQ features to obtain a third subset, wherein all candidate schemes in the third subset do not contain the new BOQ features, performing feature matching on all candidate schemes in the third subset according to the existing network BOQ features to obtain a fourth subset, calculating the feature matching degree of each candidate scheme in the fourth subset, and sequencing the candidate schemes in the fourth subset according to the feature matching degrees to obtain a candidate scheme subset. The reverse subtraction can avoid the problem of combined explosion caused by excessive scenes in forward evolution logic, and obviously reduces the number of obtained candidate scheme subsets, thereby obviously improving the speed of designing the scheme of the base station to be designed.

With reference to the second implementation manner of the first aspect of the present application, in the third implementation manner of the first aspect of the present application, the performing, according to the scheme evolution path model, recommendation degree ranking processing on all candidate schemes in the candidate scheme subset to obtain a recommendation result includes:

according to the scheme evolution path model, calculating an evolution optimization value of each candidate scheme in the candidate scheme subset through an evolution path optimization algorithm;

obtaining the recommendation degree of each candidate scheme in the candidate scheme subset according to the evolution priority value and the feature matching degree of the candidate scheme corresponding to the evolution priority value;

and sequencing all candidate schemes in the candidate scheme subset according to the recommendation degree to obtain a recommendation result.

And on the basis of a known scheme evolution path model, after processing is carried out by adopting an evolution path optimization algorithm, an evolution optimization value of each candidate scheme in the candidate scheme subset can be obtained, after the evolution optimization value of the candidate scheme is calculated, the feature matching degree of the candidate scheme is known, the recommendation degree is obtained after normalization processing is carried out on the evolution optimization value and the feature matching degree, and all the candidate schemes in the candidate scheme subset are ranked according to the recommendation degree of each candidate scheme from large to small to obtain a recommendation result.

With reference to the third embodiment of the first aspect of the present application, in the fourth embodiment of the first aspect of the present application, the method further includes:

judging whether the candidate schemes in the recommendation result belong to the second subset or the fourth subset;

if the candidate schemes in the recommendation result belong to the second subset, determining that all the candidate schemes in the recommendation result are perfect candidate schemes;

if the candidate scheme in the recommendation result belongs to the fourth subset, determining the device characteristics of the base station to be designed according to the characteristics to be used;

judging whether the candidate schemes in the recommendation result are completely matched with the device characteristics of the base station to be designed one by one;

if the first candidate scheme is not completely matched with the device characteristics of the base station to be designed, determining that the first candidate scheme is an incomplete candidate scheme;

and if the second candidate scheme is completely matched with the device characteristics of the base station to be designed, determining the second candidate scheme as a perfect candidate scheme.

Because the recommendation result is only possibly the second subset or the fourth subset, if the recommendation result corresponds to the second subset, all candidate schemes in the recommendation result are obtained under the condition that hardware is not replaceable, the device types or the number of all candidate schemes in the recommendation result are consistent with the device types or data of capacity expansion and transformation of the base station to be designed, and all candidate schemes in the recommendation result are perfect candidate schemes; if the candidate scheme in the recommendation result belongs to the fourth subset, determining the device characteristics of the base station to be designed according to the characteristics to be used, judging whether the candidate scheme in the recommendation result is completely matched with the device characteristics of the base station to be designed one by one, if the first candidate scheme is not completely matched with the device characteristics of the base station to be designed, determining that the first candidate scheme is an incomplete candidate scheme, and if the second candidate scheme is completely matched with the device characteristics of the base station to be designed, determining that the second candidate scheme is a complete candidate scheme. The perfect candidate solutions are directly recommended to the client according to the recommendation results without modification, but if the perfect candidate solutions are imperfect candidate solutions, the perfect candidate solutions cannot be directly recommended to the client, so that after the recommendation results are obtained, whether imperfect candidate solutions exist in the recommendation results or not needs to be judged.

With reference to the fourth implementation manner of the first aspect of the present application, in the fifth implementation manner of the first aspect of the present application, the performing self-perfection processing on the imperfect candidate solution according to the device connection knowledge graph to obtain a solution set of a base station to be designed includes:

determining a slot position of a port to be adjusted in the imperfect candidate scheme;

generating a frequent connecting line set through a connecting line rule mining algorithm according to the device connecting knowledge graph;

determining the setting information of the slot position of the port to be adjusted according to the frequent connection set;

adjusting the slot position of the port to be adjusted in the imperfect candidate scheme according to the setting information to obtain a self-perfecting candidate scheme;

and obtaining the base station scheme set to be designed according to the self-perfecting candidate scheme and the perfecting candidate scheme.

After an incomplete candidate scheme is determined, a port slot position to be adjusted in the incomplete candidate scheme is determined, a frequent connection set is generated through a connection rule mining algorithm under the condition that a device connection knowledge map is known, so that the frequent connection set of the port slot position to be adjusted is determined, a probability graph model (mainly applying a Bayesian network) is adopted, the matched port slot position is used as condition information, and setting information of the port slot position to be adjusted is recommended in the frequent connection set of the port slot position to be adjusted, so that base station schemes to be designed in the base station scheme set to be designed provided for a client are perfect, design and improvement are not needed, and the requirements of the client are met better.

With reference to the first aspect, the first embodiment of the first aspect, the second embodiment of the first aspect, the third embodiment of the first aspect, the fourth embodiment of the first aspect, or the fifth embodiment of the first aspect, in a sixth embodiment of the first aspect of the present application, the performing data processing on the information of the current network base station and the information of the base station to be designed to obtain a feature to be used includes:

carrying out data cleaning processing and data mapping processing on the information of the current network base station and the information of the base station to be designed;

and converting the processed information of the current network base station and the information of the base station to be designed into the characteristics to be used.

The method comprises the steps of firstly carrying out data cleaning processing on the information of the existing network base station and the information of the base station to be designed, wherein the main means of the data cleaning processing is to select effective information, eliminate useless fields (such as comments and the like), then carry out data mapping processing, for example, the type, the name and the like of a device can be analyzed from a device identification number (ID), carry out data cleaning processing and data mapping processing on the information of the existing network base station and the information of the base station to be designed, and then extract the characteristics to be used by changing the format of a file.

With reference to the sixth embodiment of the first aspect of the present application, in the seventh embodiment of the first aspect of the present application, the method further includes:

and storing the historical scheme set, the scheme evolution path model, the device connection knowledge graph and the base station scheme set to be designed to a knowledge base.

In order to enhance the performance of the base station capacity expansion modification scheme design device, the data volume of the historical scheme library needs to be enhanced, so that a more complete scheme evolution path model and device connection knowledge graph are obtained, and then the historical scheme set, the scheme evolution path model, the device connection knowledge graph and the base station scheme set to be designed can be stored in a knowledge base after being obtained each time, wherein the knowledge base is a database or a cloud storage and the like.

The second aspect of the present application provides a design apparatus for a base station capacity expansion and reconstruction scheme, including:

the acquisition module is used for acquiring the information of the current network base station, the information of the base station to be designed and a historical scheme set;

the characteristic extraction module is used for carrying out data processing on the information of the current network base station and the information of the base station to be designed to obtain the characteristics to be used;

the training learning module is used for carrying out algorithm training learning on the historical scheme set to obtain a scheme evolution path model and a device connection knowledge graph;

the scheme screening module is used for selecting a candidate scheme subset from the historical scheme set through a reverse subtraction algorithm according to the features to be used;

the scheme recommending module is used for carrying out recommendation degree ordering processing on all candidate schemes in the candidate scheme subset according to the scheme evolution path model to obtain a recommending result;

and the self-perfecting module is used for carrying out self-perfecting processing on the imperfect candidate scheme according to the device connection knowledge graph to obtain a base station scheme set to be designed when the imperfect candidate scheme exists in the recommendation result.

When a base station needs to be subjected to capacity expansion transformation, a base station design delivery person inputs current network base station information and base station information to be designed into a base station capacity expansion transformation scheme design device, an acquisition module can acquire the current network base station information and the base station information to be designed which are input by the base station design delivery person, a feature extraction module performs data processing on the current network base station information and the base station information to be designed to acquire features to be used, a historical scheme set is acquired from a network or a database of the device through the input of the base station design delivery person, a training learning module performs algorithm training learning on the historical scheme set to acquire a scheme evolution path model and a device connection knowledge map, and according to the features to be used, a scheme screening module performs feature matching on the features to be used of the historical scheme set through a reverse subtraction algorithm to select a candidate scheme subset, the scheme recommending module can deduce an evolution path of a capacity expansion transformation scheme of the base station to be designed on the basis of a scheme evolution path model, so that the recommendation degree of each candidate scheme in the candidate scheme subset is calculated, the candidate schemes in the candidate scheme subset are sequenced according to the obtained recommendation degree to obtain a recommendation result, when an incomplete candidate scheme exists in the recommendation result, the self-perfecting module can know a connection rule between devices according to a device connection knowledge map, the incomplete candidate scheme is subjected to self-perfection according to the connection rule between the devices, the incomplete candidate scheme can be completely matched with the base station to be designed, the incomplete candidate scheme after self-perfection and the original complete candidate scheme are combined into a set, and the base station to be designed scheme set is obtained. By using a reverse subtraction algorithm to obtain the candidate scheme subset, the problem of combined explosion caused by excessive scenes in forward evolution logic can be avoided, and the number of candidate schemes is obviously reduced; performing recommendation degree sorting processing on all candidate schemes in the candidate scheme subset according to the scheme evolution path model to obtain a recommendation result, so that the candidate schemes can be obtained more accurately and quickly; and performing self-perfection processing on the imperfect candidate scheme according to the device connection knowledge graph, so that the candidate scheme which better meets the requirements of customers can be provided. Therefore, the speed, the efficiency and the accuracy of the scheme design of the base station to be designed are improved.

In combination with the second aspect of the present application, in the first embodiment of the second aspect of the present application,

the training learning module is specifically used for performing evolution path training learning on the historical scheme set through a social graph mining algorithm to obtain a scheme evolution path model;

the training learning module is further used for training and learning the connection knowledge graph of the historical scheme set through a frequent item mining algorithm to obtain a device connection knowledge graph.

The training learning algorithm of the scheme evolution path model may be specifically obtained by the training learning module through a social graph mining algorithm, and the scheme evolution path model is mainly used for a process of deducing a connection relationship of devices of each capacity expansion transformation scheme in the historical scheme set. The key point of the training and learning algorithm of the device connection knowledge graph is that a training and learning module learns the basic connection rule, the deep connection relation and the network structure and parameters of a probability graph model among the devices. And obtaining a frequent association sequence and a probability graph model through a frequent item mining algorithm, and obtaining the device connection knowledge graph on the basis of the frequent association sequence and the probability graph model.

In combination with the first embodiment of the second aspect of the present application, in the second embodiment of the second aspect of the present application,

the scheme screening module is specifically used for obtaining an RF characteristic, a new BOQ characteristic, an existing network BOQ characteristic and a hardware replacement condition characteristic according to the characteristic to be used;

the scheme screening module is further configured to perform feature matching on all candidate schemes in the historical scheme set according to the RF features and the new BOQ features to obtain a first subset;

the scheme screening module is also used for judging whether the hardware is replaceable according to the hardware replacement condition characteristics;

the scheme screening module is further configured to, when the hardware replacement condition feature is not hardware replaceable, perform feature matching on all candidate schemes in the first subset according to the current network BOQ feature to obtain a second subset;

the scheme screening module is further configured to calculate a feature matching degree of each candidate scheme in the second subset, and rank the schemes in the second subset according to the feature matching degree to obtain a candidate scheme subset;

the scheme screening module is further configured to, when the hardware replacement condition feature is hardware replacement, delete and select all candidate schemes in the first subset according to the new BOQ feature to obtain a third subset, where all candidate schemes in the third subset do not include the new BOQ feature;

the scheme screening module is further configured to perform feature matching on all candidate schemes in the third subset according to the features of the existing network BOQ to obtain a fourth subset;

the scheme screening module is further configured to calculate a feature matching degree of each candidate scheme in the fourth subset, and rank the candidate schemes in the fourth subset according to the feature matching degree to obtain a candidate scheme subset.

The scheme screening module obtains RF characteristics, new BOQ characteristics, current network BOQ characteristics and hardware replacement condition characteristics according to the characteristics to be used, the RF characteristics and the new BOQ are important data for carrying out capacity expansion and reconstruction on the base station to be designed, the hardware replacement condition characteristics are preset by base station design delivery personnel and are divided into hardware replacement and hardware non-replacement, a first subset is obtained from matching the RF signature with the new BOQ signature against all of the plans in the historical plan library, judging whether the hardware is replaceable according to the hardware replacement condition characteristic, if the hardware replacement condition characteristic is not replaceable, according to the current network BOQ characteristic, performing feature matching on all candidate schemes in the first subset to obtain a second subset, calculating the feature matching degree of each candidate scheme in the second subset, sorting the schemes in the second subset according to the feature matching degree to obtain a candidate scheme subset; if the hardware replacement condition features are hardware replacement, deleting and selecting all candidate schemes in the first subset according to the new BOQ features to obtain a third subset, wherein all candidate schemes in the third subset do not contain the new BOQ features, performing feature matching on all candidate schemes in the third subset according to the existing network BOQ features to obtain a fourth subset, calculating the feature matching degree of each candidate scheme in the fourth subset, and sequencing the candidate schemes in the fourth subset according to the feature matching degrees to obtain a candidate scheme subset. The reverse subtraction can avoid the problem of combined explosion caused by excessive scenes in forward evolution logic, and obviously reduces the number of obtained candidate scheme subsets, thereby obviously improving the speed of designing the scheme of the base station to be designed.

In combination with the second embodiment of the second aspect of the present application, in the third embodiment of the second aspect of the present application,

the scheme recommending module is specifically configured to calculate, according to the scheme evolution path model, an evolution path preference algorithm to obtain an evolution preference value of each candidate scheme in the candidate scheme subset;

the scheme recommending module is further configured to obtain a recommendation degree of each candidate scheme in the candidate scheme subset according to the evolution priority value and the feature matching degree of the candidate scheme corresponding to the evolution priority value;

and the scheme recommending module is further used for sequencing all candidate schemes in the candidate scheme subset according to the recommending degree to obtain a recommending result.

On the basis of a known scheme evolution path model, after the scheme recommending module adopts an evolution path optimization algorithm for processing, an evolution optimization value of each candidate scheme in the candidate scheme subset can be obtained, after the evolution optimization value of the candidate scheme is calculated, the feature matching degree of the candidate scheme is known, the scheme recommending module obtains the recommendation degree after normalizing the evolution optimization value and the feature matching degree, and according to the recommendation degree of each candidate scheme, all candidate schemes in the candidate scheme subset are ranked in a mode from large to small to obtain a recommendation result.

In combination with the third embodiment of the second aspect of the present application, in the fourth embodiment of the second aspect of the present application,

the self-perfecting module is further configured to determine that the candidate solution in the recommendation result belongs to the second subset or the fourth subset;

the self-perfecting module is further configured to determine that all candidate schemes in the recommendation result are perfect candidate schemes when the candidate schemes in the recommendation result belong to the second subset;

the self-perfecting module is further configured to determine, when the candidate solution in the recommendation result belongs to the fourth subset, a device feature of the base station to be designed according to the feature to be used;

the self-perfecting module is further used for judging whether the candidate schemes in the recommendation result are completely matched with the device characteristics of the base station to be designed one by one;

the self-perfecting module is further configured to determine that the first candidate scheme is an imperfect candidate scheme when the first candidate scheme is not completely matched with the device characteristics of the base station to be designed;

the self-perfecting module is further configured to determine that the second candidate scheme is a perfect candidate scheme when the second candidate scheme is completely matched with the device characteristics of the base station to be designed.

Because the recommendation result is only possibly the second subset or the fourth subset, if the recommendation result corresponds to the second subset, all candidate schemes in the recommendation result are obtained under the condition that hardware is not replaceable, the device types or the number of all candidate schemes in the recommendation result are consistent with the device types or data of capacity expansion and transformation of the base station to be designed, and all candidate schemes in the recommendation result are perfect candidate schemes; if the candidate scheme in the recommendation result belongs to the fourth subset, the self-perfecting module determines the device characteristics of the base station to be designed according to the characteristics to be used, judges whether the candidate scheme in the recommendation result is completely matched with the device characteristics of the base station to be designed one by one, if the first candidate scheme is not completely matched with the device characteristics of the base station to be designed, the self-perfecting module determines that the first candidate scheme is an incomplete candidate scheme, and if the second candidate scheme is completely matched with the device characteristics of the base station to be designed, the second candidate scheme is a perfect candidate scheme. The perfect candidate solutions are directly recommended to the client according to the recommendation results without modification, but if the perfect candidate solutions are imperfect candidate solutions, the perfect candidate solutions cannot be directly recommended to the client, so that after the recommendation results are obtained, whether imperfect candidate solutions exist in the recommendation results or not needs to be judged.

In combination with the fourth embodiment of the second aspect of the present application, in the fifth embodiment of the second aspect of the present application,

the self-perfecting module is further used for determining a slot position of a port to be adjusted in the imperfect candidate scheme;

the self-improvement module is also used for generating a frequent connecting line set through a connecting line rule mining algorithm according to the device connecting knowledge graph;

the self-perfecting module is also used for determining the setting information of the slot position of the port to be adjusted according to the frequent connection set;

the self-perfecting module is also used for adjusting the slot position of the port to be adjusted in the imperfect candidate scheme according to the setting information to obtain a self-perfecting candidate scheme;

and the self-perfecting module is also used for obtaining the base station scheme set to be designed according to the self-perfecting candidate scheme and the perfecting candidate scheme.

After the incomplete candidate scheme is determined by the self-perfecting module, a slot position of a port to be adjusted in the incomplete candidate scheme is determined, under the condition that a device connection knowledge map is known, a frequent connection set is generated through a connection rule mining algorithm, so that the frequent connection set of the slot position of the port to be adjusted is determined, a probability graph model (mainly applying a Bayesian network) is adopted, the matched port slot position is used as condition information, and setting information of the slot position of the port to be adjusted is recommended in the frequent connection set of the slot position of the port to be adjusted, so that base station schemes to be designed in the base station scheme set to be designed, which are provided for clients, are perfect, do not need to be designed and improved, and meet the requirements of the clients better.

With reference to the second aspect of the present application, the first embodiment of the second aspect, the second embodiment of the second aspect, the third embodiment of the second aspect, the fourth embodiment of the second aspect, or the fifth embodiment of the second aspect, in the sixth embodiment of the second aspect of the present application,

the characteristic extraction module is specifically used for carrying out data cleaning processing and data mapping processing on the existing network base station information and the base station information to be designed;

the feature extraction module is further configured to convert the processed current network base station information and the information of the base station to be designed into features to be used.

The feature extraction module firstly performs data cleaning processing on the information of the current network base station and the information of the base station to be designed, the data cleaning processing mainly comprises the steps of selecting effective information, eliminating useless fields (such as comments and the like), then performing data mapping processing, for example, the type, the name and the like of a device can be analyzed from a device identification number (ID), performing data cleaning processing and data mapping processing on the information of the current network base station and the information of the base station to be designed, and extracting features to be used by the feature extraction module through changing a file format.

In combination with the sixth embodiment of the second aspect of the present application, in the seventh embodiment of the second aspect of the present application, the apparatus further comprises:

and the knowledge base storage module is used for storing the historical scheme set, the scheme evolution path model, the device connection knowledge graph and the base station scheme set to be designed into a knowledge base.

In order to enhance the performance of the base station capacity expansion modification scheme design device, the data volume of the historical scheme library needs to be enhanced, so that a more complete scheme evolution path model and device connection knowledge graph are obtained, and then after the historical scheme set, the scheme evolution path model, the device connection knowledge graph and the base station scheme set to be designed are obtained each time, the knowledge library storage module can store the historical scheme set, the scheme evolution path model, the device connection knowledge graph and the base station scheme set to be designed into a knowledge library, wherein the knowledge library is a database or a cloud storage and the like.

A third aspect of the present application provides a server comprising:

a processor, a transceiver, and a memory, wherein the memory may be used to store code executed by the processor;

the processor, the transceiver and the memory are connected through a bus system;

the transceiver is used for acquiring the information of the current network base station, the information of the base station to be designed and a historical scheme set;

the processor is used for carrying out data processing on the information of the current network base station and the information of the base station to be designed to obtain the characteristics to be used;

the processor is used for carrying out algorithm training learning on the historical scheme set to obtain a scheme evolution path model and a device connection knowledge map;

the processor is used for selecting a candidate scheme subset from the historical scheme set through a reverse subtraction algorithm according to the features to be used;

the processor is used for carrying out recommendation degree sorting processing on all candidate schemes in the candidate scheme subset according to the scheme evolution path model to obtain a recommendation result;

and the processor is used for performing self-perfection processing on the imperfect candidate scheme according to the device connection knowledge graph to obtain a base station scheme set to be designed when the imperfect candidate scheme exists in the recommendation result.

When a base station needs to be subjected to capacity expansion transformation, a base station design delivery person inputs the information of the current network base station and the information of the base station to be designed into a base station capacity expansion transformation scheme design device, a transceiver can acquire the information of the current network base station and the information of the base station to be designed, a processor performs data processing on the information of the current network base station and the information of the base station to be designed to acquire a feature to be used, and acquires a historical scheme set from a network or a database of the device through the input of the base station design delivery person, the processor performs algorithm training and learning on the historical scheme set to acquire a scheme evolution path model and a device connection knowledge map, and performs feature matching on the historical scheme set through a reverse subtraction algorithm according to the feature to be used so as to select a candidate scheme subset, and the processor is based on the scheme evolution path model, the evolution path of the capacity expansion transformation scheme of the base station to be designed can be deduced, so that the recommendation degree of each candidate scheme in the candidate scheme subset is calculated, the candidate schemes in the candidate scheme subset are sequenced according to the obtained recommendation degree, a recommendation result is obtained, when an incomplete candidate scheme exists in the recommendation result, the processor can know the connection rule between the devices according to the device connection knowledge map, the incomplete candidate scheme is subjected to self-perfection processing according to the connection rule between the devices, the incomplete candidate scheme can be completely matched with the base station to be designed, the incomplete candidate scheme after the self-perfection processing and the originally complete candidate scheme are combined into a set, and the base station scheme set to be designed is obtained. By using a reverse subtraction algorithm to obtain the candidate scheme subset, the problem of combined explosion caused by excessive scenes in forward evolution logic can be avoided, and the number of candidate schemes is obviously reduced; performing recommendation degree sorting processing on all candidate schemes in the candidate scheme subset according to the scheme evolution path model to obtain a recommendation result, so that the candidate schemes can be obtained more accurately and quickly; and performing self-perfection processing on the imperfect candidate scheme according to the device connection knowledge graph, so that the candidate scheme which better meets the requirements of customers can be provided. Therefore, the speed, the efficiency and the accuracy of the scheme design of the base station to be designed are improved.

A fourth aspect of the present application provides a computer-readable storage medium, which includes instructions that, when executed on a computer, cause the computer to execute the above method for designing a base station capacity expansion modification scheme.

A fifth aspect of the present application provides a computer program product containing instructions, which when run on a computer, causes the computer to execute the above method for designing a base station capacity expansion modification scheme.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following briefly introduces the embodiments and the drawings used in the description of the prior art, and obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without inventive efforts.

Fig. 1 is a schematic view of an application scenario provided in the present application;

fig. 2 is a schematic flow chart of a method for designing a base station capacity expansion and transformation scheme according to the present application;

FIG. 3 is a schematic diagram of a device frequent association sequence provided herein;

FIG. 4 is a schematic diagram of a probabilistic graphical model provided herein;

FIG. 5 is a schematic flow chart of an inverse subtraction algorithm provided herein;

fig. 6 is an example of xml format of the current network base station information and the base station information to be designed provided by the present application;

FIG. 7 is an example of a JSON format of a feature expression provided herein;

fig. 8 is a schematic structural diagram of a device for designing a base station capacity expansion and transformation scheme according to the present application;

fig. 9 is a schematic structural diagram of another apparatus for designing a base station capacity expansion and transformation scheme provided in the present application;

fig. 10 is a schematic structural diagram of a server provided in the present application.

Detailed Description

The technical solutions in the present application will be described clearly and completely with reference to the accompanying drawings in the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

First, a system architecture or scenario in which the present application is applied will be briefly described.

The deployment scheme under the current base station main equipment capacity expansion and reconstruction scene needs to be designed based on the existing network stock, the design process is completed manually, and the difference lies in a graph showing mode.

And in a drawing mode I, PPT is manufactured for pure manual work. The method specifically comprises the following steps: based on customer requirements, BOQ and current network inventory configuration data, the PPT is manually made based on expert experience.

And the drawing mode II is used for automatically drawing the tool. The method specifically comprises the following steps: importing existing network stock configuration data, and finishing the restoration of an existing network base station deployment scheme by using a tool; based on the current network scheme and the configuration information of products before sale, and depending on expert experience, a target network scheme meeting RF planning is designed manually; and filling and importing a main device wiring rule table and a site information table (BOQ and RF planning information) based on the design result, reading parameters by the tool, and automatically generating a main device planning diagram based on the built-in hardware rule.

The following disadvantages exist in the above first and second drawing modes: 1. the capacity expansion transformation is carried out in a forward mode, in the scheme design, the types of the devices are more, the quantity selection is not fixed, the slot position selection of the devices and the port connection mode selection between the devices are more, the combination explosion occurs, namely the number of the permutation and combination is too large, the quantity of the candidate schemes which can be generated is huge, and the final scheme is difficult to judge and select by manpower from many combination designs; 2. the design of the deployment scheme of the capacity expansion reconstruction scene needs to refer to the existing network scheme, the design process is complex, and the experience requirement on technicians is higher than that of a newly-built scene; 3. the design link is completely completed by depending on expert experience, so that the time is consumed and the error rate is high; 4. the current scheme does not realize cloud storage, and the optimal scheme cannot be shared on the global scale. In the above four points, the speed and efficiency of designing the capacity expansion and transformation scheme of the base station are low.

In order to improve the speed and efficiency of designing the base station capacity expansion transformation scheme, the application provides a base station capacity expansion transformation scheme designing method and a related device, which are used for improving the speed and efficiency of designing the base station capacity expansion transformation scheme, an applied system architecture or a scene is shown in fig. 1. It should be noted that the method of the present application is not limited to the design of the base station capacity expansion transformation scheme, and may also be applied to other equipment occasions where the scene capacity expansion transformation needs to be performed, including application to other industrial drawing design services, such as buildings, power, machinery, and the like.

As shown in fig. 2, an embodiment of the present application provides a method for designing a capacity expansion and reconstruction scheme of a base station, including:

201. acquiring information of a current network base station, information of a base station to be designed and a historical scheme set;

in this embodiment, when a base station needs to be subjected to capacity expansion transformation, a base station design delivery person inputs information of an existing network base station and information of a base station to be designed into a base station capacity expansion transformation scheme design device, and the base station capacity expansion transformation scheme design device can acquire the information of the existing network base station and the information of the base station to be designed, which are input by the base station design delivery person, and acquire a history scheme set from a network or a database of the device, where the history scheme set is a capacity expansion transformation scheme during history capacity expansion transformation of the existing base station.

202. Carrying out data processing on the information of the current network base station and the information of the base station to be designed to obtain the characteristics to be used;

in this embodiment, data processing is performed on the acquired information of the existing network base station and the information of the base station to be designed, including data cleaning, feature extraction, and the like, to obtain features to be used, where the features to be used generally include device features, and the device features include device types, device numbers, device slots, device ports, device connection relationships, and the like.

203. Carrying out algorithm training learning on the historical scheme set to obtain a scheme evolution path model and a device connection knowledge map;

in this embodiment, the historical solution set is subjected to algorithm training and learning to obtain a solution evolution path model and a device connection knowledge graph, and since the base station expansion modification process is not one-kick-on but a process of gradually adding/replacing devices and adding/modifying connection relationships between the devices, a process of obtaining connection relationships between the devices of each expansion modification solution in the historical solution set can be derived from the solution evolution path model, and connection rules between the devices can be known from the device connection knowledge graph, and the connection rules between the devices specifically include connection rules between port slots of the devices, and the like.

204. Selecting a candidate scheme subset from the historical scheme set through a reverse subtraction algorithm according to the features to be used;

in this embodiment, according to the feature to be used, the feature matching of the feature to be used is performed on the history scheme set through a reverse subtraction algorithm, so that the candidate scheme subset is selected.

205. Performing recommendation degree sorting processing on all candidate schemes in the candidate scheme subset according to the scheme evolution path model to obtain a recommendation result;

in this embodiment, on the basis of the scheme evolution path model obtained in step 203, an evolution path of the capacity expansion transformation scheme of the base station to be designed may be derived, so as to calculate the recommendation degree of each candidate scheme in the candidate scheme subset, and rank the candidate schemes in the candidate scheme subset according to the obtained recommendation degree, so as to obtain a recommendation result.

206. And when the incomplete candidate scheme exists in the recommendation result, performing self-perfection processing on the incomplete candidate scheme according to the device connection knowledge graph to obtain a base station scheme set to be designed.

In this embodiment, if the candidate solution in the recommendation result may have a condition that the types or the numbers of the devices are not consistent with the base station to be designed, the connection rule between the devices of the candidate solution may not match the actual device of the base station to be designed, such a candidate solution is referred to as an incomplete candidate solution, at this time, the connection rule between the devices can be known according to the device connection knowledge map, and the incomplete candidate solution is self-perfected according to the connection rule between the devices, so that the incomplete candidate solution can be completely matched with the base station to be designed, and the incomplete candidate solution after self-perfection and the originally perfect candidate solution are merged into one set to obtain a base station to be designed solution set.

In the embodiment of the application, the candidate scheme subset is obtained by using a reverse subtraction algorithm, so that the problem of combined explosion caused by excessive scenes in forward evolution logic can be avoided, and the number of candidate schemes is obviously reduced; performing recommendation degree sorting processing on all candidate schemes in the candidate scheme subset according to the scheme evolution path model to obtain a recommendation result, so that the candidate schemes can be obtained more accurately and quickly; and performing self-perfection processing on the imperfect candidate scheme according to the device connection knowledge graph, so that the candidate scheme which better meets the requirements of customers can be provided. Therefore, the speed, the efficiency and the accuracy of the scheme design of the base station to be designed are improved.

Optionally, in some embodiments of the present application, performing algorithm training and learning on the historical solution set to obtain an evolution knowledge graph and a device connection knowledge graph includes:

In the embodiment of the present application, the training learning algorithm of the scheme evolution path model may be obtained by a social graph mining algorithm, and the scheme evolution path model is mainly used for deducing a connection relationship between devices of each capacity expansion transformation scheme in the historical scheme set, and an evolution path optimization algorithm is adopted based on the scheme evolution path model, the evolution path optimization algorithm is a specific problem of the maximum flow and minimum cost flow problems of the network, and there are many classical algorithms and improved algorithms thereof regarding the maximum flow and minimum cost flow problems of the network, such as Ford-Fulkerson algorithm, Dijkstra algorithm, negative loop algorithm, and the like. The social graph mining algorithm for obtaining the scheme evolution path model mainly adopts an analysis method based on graph theory to represent the whole social network in a graph form, wherein the nodes represent individuals in the social network, and the edges represent connections among the individuals. The research results include small world theory, centrality analysis, community analysis and the like. The method is applied to the expansion transformation of base station main equipment, and uses nodes to represent devices and edges to represent connecting lines among the devices;

the key points of the training and learning algorithm of the device connection knowledge graph are to learn the basic connection rules and deep connection relations among the devices and the network structure and parameters of the probability graph model. As shown in fig. 3, a schematic diagram of a frequent association sequence of devices is shown, ABCD respectively represents different device and port information, such as a: BBU _1-Port0, B: RXU _2-Port1, AB indicates that Port0 of BBU _1 is connected to Port1 of RXU _2, and so on. The frequent item mining algorithm may be an Association rule mining algorithm (Association rule mining), which is one of the most active research methods in data mining, and may be used to find the relation between things, at the earliest, in order to find the relation between different commodities in a supermarket transaction database, and the most classical case is diapers and beer in walmart. Common algorithms include Apriori algorithm, FP-Growth algorithm, and the like. The FP-grow algorithm is a correlation analysis algorithm proposed by korean weir et al in 2000, and compresses a database providing frequent item sets to a frequent pattern tree (FP-tree), but still retains item set correlation information. The association rule mining is mainly used for learning basic connection rules among devices, for example, BBU _1 can be connected with RXU _2, BBU _1 cannot be connected with a cabin, and the like;

in the process of expanding and transforming the base station, the addition and modification of the connecting lines among the devices are involved, and under the condition, a plurality of redundant connecting lines can be obtained only according to the basic connecting rules among the devices, so that the effectiveness of the scheme is reduced. The probabilistic graphical model shown in fig. 4 is a method for learning deep connectivity effectively, in which A, B, C and D sub-tables represent port information of different devices, such as a: BBU _1-Port0, B: RXU _2-Port1, C: BBU _2-Port1, D: RRU _1-Port 2. P (AD) represents the probability of A being connected to D, and P (AB | BC) represents the probability of A being connected to B under the condition that B is connected to C. Probabilistic graphical models can be roughly divided into two categories: bayesian networks (Bayesian networks) and Markov Random fields (Markov Random fields). The main differences are that different types of graphs are used to express relationships between variables, the bayesian network uses directed acyclic graphs (DirectedAcyclic graphs) to express causal relationships, and the markov random field uses undirected graphs (undirected graphs) to express interactions between variables. Because of the difficulty of parameter estimation for markov fields, we mainly use bayesian networks. Each node in the bayesian network corresponds to a prior probability distribution or a conditional probability distribution, so that the overall joint distribution can be directly decomposed into the product of the distributions corresponding to all the individual nodes. The deep connection relation can give the probability that A is connected with B under a certain condition, for example, under the condition that B is connected with C in fig. 4, namely, the conditional probability, so that the conditional probability distribution can be obtained by other information, the deep connection relation is obtained by obtaining the joint distribution, and the network structure and the parameters of the probability graph model are trained according to the historical scheme, wherein the network structure is a device connection relation graph, and the parameters are the probability that devices are connected with edges. Then a device connection knowledge map can be obtained based on the frequent association sequence and the probabilistic map model.

Optionally, in some embodiments of the present application, selecting a candidate scheme subset from the historical scheme set by using a reverse subtraction algorithm according to the feature to be used includes:

obtaining an RF characteristic, a new BOQ characteristic, an existing network BOQ characteristic and a hardware replacement condition characteristic according to the characteristic to be used;

performing feature matching on all candidate schemes in the third subset according to the characteristics of the current network BOQ to obtain a fourth subset;

and calculating the feature matching degree of each candidate scheme in the fourth subset, and sequencing the candidate schemes in the fourth subset according to the feature matching degree to obtain the candidate scheme subset.

In this embodiment of the present application, it is assumed that information of a historical solution library is represented in a three-layer aggregate manner, a first-layer aggregate represents the entire solution library, a second-layer aggregate represents certain base station solution information in the solution library, and a third-layer aggregate represents three different types of information included in each base station solution, so that when a base station to be designed is subjected to capacity expansion transformation, a three-layer aggregate manner is adopted, and a specific process of selecting a candidate solution subset from the historical solution library through a reverse subtraction algorithm is as follows:

(1) obtaining an RF characteristic, a new BOQ characteristic, an existing network BOQ characteristic and a hardware replacement condition characteristic according to the characteristic to be used;

the RF characteristic represents the capacity of generating high-frequency alternating current change electromagnetic waves and is important data for carrying out capacity expansion transformation on a base station to be designed, the new BOQ characteristic represents partial project items, measure projects, other project names and a detailed list with a corresponding quantity of a proposed project of the base station to be designed after capacity expansion transformation, the existing network BOQ characteristic represents partial project items, measure projects, other project names and a detailed list with a corresponding quantity of the proposed project of the existing network base station after capacity expansion transformation, and the hardware replacement condition characteristic is preset by a base station design delivery personnel and is divided into hardware replacement and hardware irreplaceable.

(2) Performing feature matching on all candidate schemes in the historical scheme set according to the RF features and the new BOQ features to obtain a first subset;

referring to fig. 5, a first subset, i.e., a set S, is obtained by matching RF characteristics and new BOQ characteristics (where RF and new BOQ belong to a third-level set and belong to different subsets) of the capacity expansion modified base station with all solutions in the historical solution library U.

(3) Judging whether the hardware is replaceable according to the hardware replacement condition characteristics;

(4) if the hardware replacement condition characteristic is not hardware replaceable, performing characteristic matching on all candidate schemes in the first subset according to the current network BOQ characteristic to obtain a second subset;

if all hardware is not replaceable, it is simpler to match the solution in the first subset (S) directly with the features of the existing net BOQ, resulting in a second subset (a) that meets the requirements.

(5) Calculating the feature matching degree of each candidate scheme in the second subset, and sequencing the schemes in the second subset according to the feature matching degree to obtain a candidate scheme subset;

and (4) calculating the feature matching degree d of the second subset (A) and the base station to be reconstructed according to the other features such as the port slot position and the like, and then sequencing according to the matching degree from high to low to obtain the candidate scheme subset. The calculation formula of the feature matching degree d is as follows:

wherein, subscript 1 of d refers to the base station to be modified, and 2 refers to the 2 nd scheme in the second subset (a); x is the number of_1kNamely, the kth item in the n items of characteristics of the base station to be reconstructed refers to one of the characteristics of a port, a slot position and the like; x is the number of_2kThat is, the kth item of each of the n items of the 2 nd scheme in the second subset (a) refers to one of the characteristics of a port, a slot, and the like.

(6) If the hardware replacement condition characteristic is hardware replacement, deleting all candidate schemes in the first subset according to the new BOQ characteristic to obtain a third subset, wherein all candidate schemes in the third subset do not contain the new BOQ characteristic;

if the hardware is replaceable, the scheme matching the new BOQ feature is subtracted from the first subset (S) to obtain a third subset (B).

(7) Performing feature matching on all candidate schemes in the third subset according to the features of the current network BOQ to obtain a fourth subset;

matching the current net BOQ features with the schemes in the third subset (B) to obtain a fourth subset (C) which meets the requirements.

(8) And calculating the feature matching degree of each candidate scheme in the fourth subset, and sequencing the candidate schemes in the fourth subset according to the feature matching degree to obtain the candidate scheme subset.

The specific way of calculating the feature matching degree is the same as that in the step (5), and details are not repeated.

As can be seen from (1) to (8) above, the reverse subtraction can avoid the problem of combinatorial explosion caused by too many scenes in the forward evolution logic, and the number of the obtained candidate scheme subsets is obviously reduced, so that the speed of designing the scheme of the base station to be designed can be obviously increased.

Optionally, in some embodiments of the present application, performing recommendation degree ranking processing on all candidate solutions in the candidate solution subset according to the solution evolution path model to obtain a recommendation result includes:

In the embodiment of the present application, after the feature matching degree of each candidate scheme in the candidate scheme subset is obtained in (5) or (8) described above, and on the basis of the known scheme evolution path model, after processing is performed by using the evolution path preference algorithm, an evolution preference value of each candidate scheme in the candidate scheme subset can be obtained, and a specific calculation process is as follows: aiming at project/region/country level, calculating the preferred value of each candidate scheme in the candidate scheme subset according to the capacity expansion transformation/evolution path among the candidate schemes, taking a webpage ranking (PageRank) as an example, the PageRank is a technology calculated according to mutual hyperlinks among webpages, is one of elements of the webpage ranking, is used for embodying the relevance and the importance of the webpages, is applied to the capacity expansion transformation of base station main equipment, and embodies the relevance and the importance of devices by the PageRank value. The evolution optimization value is obtained by calculating a scheme evolution path, and the formula is as follows:

in the above formula, PR (p)_i) Representing devices p in candidate i_iA PageRank value of (a), i.e., an evolution preference value; p is a radical of_i、p_jAre representative of different devices, p_jIs a device in candidate j; d is a coefficient greater than 0 and less than 1; l (j) represents p_jNumber of devices connected, p since d is preset_i、p_jAnd L (j) can be obtained by a scheme evolution path model, then the evolution preference value PR (p)_i) Can be calculated;

preferred value of evolution PR (p) at candidate i_i) After calculation, the feature matching degree d of the candidate scheme i is known_iTo evolve a preferred value PR (p)_i) Degree of matching with features d_iCarrying out normalization processing, wherein the normalization formula is as follows:

R_i＝(1-α)PR_i+α*d_i

wherein α is a constant greater than 0 and less than 1, the recommendation degree R of the candidate i can be obtained_i；

And sequencing all the candidate schemes in the candidate scheme subset from large to small according to the recommendation degree of each candidate scheme to obtain a recommendation result.

Alternatively, in some embodiments of the present application,

In this embodiment of the application, since the recommendation result may only correspond to the second subset in the above embodiment (4) or the fourth subset in (7), if the recommendation result corresponds to the second subset, it indicates that all candidate schemes in the recommendation result are obtained under the condition that hardware is not replaceable, and then it indicates that the device types or the number of all candidate schemes in the recommendation result are consistent with the device types or data of capacity expansion and transformation of the base station to be designed, and all candidate schemes in the recommendation result are perfect candidate schemes; if the candidate scheme in the recommendation result belongs to the fourth subset, determining the device characteristics of the base station to be designed according to the characteristics to be used, judging whether the candidate scheme in the recommendation result is completely matched with the device characteristics of the base station to be designed one by one, if the first candidate scheme is not completely matched with the device characteristics of the base station to be designed, determining that the first candidate scheme is an incomplete candidate scheme, and if the second candidate scheme is completely matched with the device characteristics of the base station to be designed, determining that the second candidate scheme is a complete candidate scheme. The perfect candidate solutions are directly recommended to the client according to the recommendation results without modification, but if the perfect candidate solutions are imperfect candidate solutions, the perfect candidate solutions cannot be directly recommended to the client, so that after the recommendation results are obtained, whether imperfect candidate solutions exist in the recommendation results or not needs to be judged.

Optionally, in some embodiments of the present application, performing self-improvement processing on an incomplete candidate solution according to a device connection knowledge graph to obtain a solution set of a base station to be designed, where the method includes:

determining a slot position of a port to be adjusted in an imperfect candidate scheme;

generating a frequent connection set through a connection rule mining algorithm according to the device connection knowledge graph;

and obtaining a base station scheme set to be designed according to the self-perfecting candidate scheme and the perfecting candidate scheme.

In the embodiment of the present application, after an incomplete candidate scheme is determined, a slot position of a port to be adjusted in the incomplete candidate scheme is determined, and under the condition that a device connection knowledge map is known, a frequent connection set is generated by a connection rule mining algorithm, so that a frequent connection set of the slot position of the port to be adjusted is determined, and then a probability graph model (mainly using a bayesian network) is adopted, the matched port slot position is used as condition information, and a value of the slot position of the port to be adjusted is recommended in the frequent connection set of the slot position of the port to be adjusted, as shown in table 1 and table 2, a table sample is arranged for port connection and slot position of the candidate scheme that is finally recommended, and a connection relationship between ports (Self-port 0 and 1) and ports (Head-type-port) on a single board (u _1) is given in table 1, that is a port0 on the single board RXU _1 is connected with BBU _1 _ port2, a port _2, The recommendation degrees of the three ports, namely the BBU _2port0 and the BBU _3port4, are 0.7, 0.2 and 0.1, respectively, according to the maximum recommendation degree principle, the port0 port of the board RXU _1 is connected to the port2 of the board BBU _1 in the final scheme, and so on. Table 2 shows the recommendation degrees of arrangement of the single board RXU _1 in the slot positions 0, 1, 2, and 3, which are 0.5, 0.3, 0.1, and 0.1, respectively, and according to the maximum recommendation degree principle, the RXU _1 in the final scheme is placed in the slot position 0. Therefore, when an imperfect candidate scheme exists, the setting information of the slot position of the port to be adjusted can be given, so that the base station schemes to be designed in the base station scheme set to be designed provided for the client are perfect, the design and the improvement are not needed, and the requirements of the client are met better.

TABLE 1

TABLE 2

Optionally, in some embodiments of the present application, data processing is performed on information of a current network base station and information of a base station to be designed to obtain a feature to be used, where the feature includes:

In the embodiment of the application, data cleaning processing is performed on existing network base station information and base station information to be designed, the main means of the data cleaning processing is to select effective information, remove useless fields (such as comments and the like), perform data mapping processing, for example, the device type and name can be analyzed from a device identification number (ID), perform data cleaning processing and data mapping processing on the existing network base station information and the base station information to be designed, and extract features to be used by changing the format of a file, as shown in fig. 6, the features are examples of the existing network base station information and the base station information to be designed in an xml format, and fig. 7 is an example of a JSON format converted into a feature expression. For example, BBU type1 in fig. 7 represents a specific type of BBU, and other device type expressions such as BBU _ type2, …, RXU _ type1, AAU _ type1, Antenna _ type1, TMA _ type1, Combiner _ type1, RCU _ type1, SBT _ type1, splitter _ type1, and the like, and feature data such as the number of devices, device slot (sector no), device ports (ports), and device wiring relationship. An example of a characterization is shown in fig. 7.

Optionally, in some embodiments of the present application, the method further includes:

and storing the historical scheme set, the scheme evolution path model, the device connection knowledge graph and the base station scheme set to be designed into a knowledge base.

In the embodiment of the application, in order to enhance the performance of the base station capacity expansion and transformation scheme design device, the data size of the historical scheme library needs to be enhanced, so that a more complete scheme evolution path model and a device connection knowledge graph are obtained, and then the historical scheme set, the scheme evolution path model, the device connection knowledge graph and the base station scheme set to be designed can be stored in a knowledge base after being obtained each time, wherein the knowledge base is a database or a cloud storage and the like.

The above embodiments describe a method for designing a base station capacity expansion and reconstruction scheme, and a device for designing a base station capacity expansion and reconstruction scheme is described below with reference to the embodiments.

As shown in fig. 8, an embodiment of the present application provides a device for designing a capacity expansion and reconstruction scheme of a base station, including:

an obtaining module 801, configured to obtain information of a current network base station, information of a base station to be designed, and a history scheme set;

the feature extraction module 802 is configured to perform data processing on the information of the current network base station and the information of the base station to be designed to obtain a feature to be used;

a training learning module 803, configured to perform algorithm training learning on the historical scheme set to obtain a scheme evolution path model and a device connection knowledge graph;

the scheme screening module 804 is used for selecting a candidate scheme subset from the historical scheme set through a reverse subtraction algorithm according to the features to be used;

the scheme recommending module 805 is configured to perform recommendation degree sorting processing on all candidate schemes in the candidate scheme subset according to the scheme evolution path model to obtain a recommendation result;

and a self-perfecting module 806, configured to, when an incomplete candidate solution exists in the recommendation result, perform self-perfecting processing on the incomplete candidate solution according to the device connection knowledge graph, to obtain a to-be-designed base station solution set.

In the embodiment of the application, when a base station needs to be subjected to capacity expansion transformation, a base station design delivery person inputs current network base station information and information of a base station to be designed into a base station capacity expansion transformation scheme design device, an acquisition module 801 can acquire the current network base station information and the information of the base station to be designed input by the base station design delivery person, a feature extraction module 802 performs data processing on the current network base station information and the information of the base station to be designed to obtain a feature to be used, and acquires a historical scheme set from a network or a database of the device through the input of the base station design delivery person, a training learning module 803 performs algorithm training learning on the historical scheme set to obtain a scheme evolution path model and a device connection knowledge map, and a scheme screening module 804 performs feature matching on the historical scheme set through a reverse subtraction algorithm according to-be-used feature, thereby selecting a candidate scheme subset, the scheme recommending module 805 can deduce an evolution path of the capacity-expanding transformation scheme of the base station to be designed on the basis of the scheme evolution path model, thereby calculating the recommendation degree of each candidate scheme in the candidate scheme subset, sorting the candidate schemes in the candidate scheme subset according to the obtained recommendation degree to obtain a recommendation result, when an imperfect candidate scheme exists in the recommendation result, the self-perfecting module 806 can know the connection rule between the devices according to the device connection knowledge graph, and realizes the self-perfecting processing on the imperfect candidate scheme according to the connection rule between the devices, so that the imperfect candidate scheme can completely match the base station to be designed, and the imperfect candidate scheme after the self-perfecting processing and the originally perfect candidate scheme are combined into a set, and obtaining a base station scheme set to be designed. By using a reverse subtraction algorithm to obtain the candidate scheme subset, the problem of combined explosion caused by excessive scenes in forward evolution logic can be avoided, and the number of candidate schemes is obviously reduced; performing recommendation degree sorting processing on all candidate schemes in the candidate scheme subset according to the scheme evolution path model to obtain a recommendation result, so that the candidate schemes can be obtained more accurately and quickly; and performing self-perfection processing on the imperfect candidate scheme according to the device connection knowledge graph, so that the candidate scheme which better meets the requirements of customers can be provided. Therefore, the speed, the efficiency and the accuracy of the scheme design of the base station to be designed are improved.

Alternatively, in some embodiments of the present application,

a training learning module 803, specifically configured to perform evolution path training learning on the historical scheme set through a social graph mining algorithm to obtain a scheme evolution path model;

the training learning module 803 is further configured to perform connection knowledge graph training learning on the historical scheme set through a frequent item mining algorithm to obtain a device connection knowledge graph.

In this embodiment of the application, the training learning algorithm of the scheme evolution path model may be specifically obtained by the training learning module 803 through a social graph mining algorithm, and the scheme evolution path model is mainly used for a process of deducing a connection relationship between devices that may obtain each capacity expansion transformation scheme in the historical scheme set. The key point of the training and learning algorithm of the device connection knowledge graph is that the training and learning module 803 learns the basic connection rules, the deep connection relations and the network structure and parameters of the probability graph model among the devices. And obtaining a frequent association sequence and a probability graph model through a frequent item mining algorithm, and obtaining the device connection knowledge graph on the basis of the frequent association sequence and the probability graph model.

Alternatively, in some embodiments of the present application,

the scheme screening module 804 is specifically used for obtaining an RF characteristic, a new BOQ characteristic, an existing network BOQ characteristic and a hardware replacement condition characteristic according to the characteristic to be used;

the scheme screening module 804 is further configured to perform feature matching on all candidate schemes in the historical scheme set according to the RF features and the new BOQ features to obtain a first subset;

the scheme screening module 804 is further configured to determine whether the hardware is replaceable according to the hardware replacement condition characteristic;

the scheme screening module 804 is further configured to, when the hardware replacement condition feature is not hardware replaceable, perform feature matching on all candidate schemes in the first subset according to a feature of a current network BOQ to obtain a second subset;

the scheme screening module 804 is further configured to calculate a feature matching degree of each candidate scheme in the second subset, and rank the schemes in the second subset according to the feature matching degree to obtain a candidate scheme subset;

the scheme screening module 804 is further configured to, when the hardware replacement condition feature is hardware replacement, delete and select all candidate schemes in the first subset according to the new BOQ feature to obtain a third subset, where all candidate schemes in the third subset do not include the new BOQ feature;

the scheme screening module 804 is further configured to perform feature matching on all candidate schemes in the third subset according to features of the existing network BOQ to obtain a fourth subset;

the scheme screening module 804 is further configured to calculate a feature matching degree of each candidate scheme in the fourth subset, and rank the candidate schemes in the fourth subset according to the feature matching degree to obtain a candidate scheme subset.

In this embodiment, the scheme screening module 804 obtains an RF feature, a new BOQ feature, an existing network BOQ feature, and a hardware replacement condition feature according to a feature to be used, where the RF feature and the new BOQ are important data for capacity expansion and transformation of a base station to be designed, the hardware replacement condition feature is preset by a base station design delivery person and is divided into a hardware replaceable feature and a hardware non-replaceable feature, matches the RF feature and the new BOQ feature with all schemes in a historical scheme library to obtain a first subset, determines whether the hardware is replaceable according to the hardware replacement condition feature, if the hardware replacement condition feature is not replaceable, performs feature matching on all candidate schemes in the first subset according to the existing network BOQ feature to obtain a second subset, calculates a feature matching degree of each candidate scheme in the second subset, sorts the schemes in the second subset according to the feature matching degree, obtaining a candidate scheme subset; if the hardware replacement condition features are hardware replacement, deleting and selecting all candidate schemes in the first subset according to the new BOQ features to obtain a third subset, wherein all candidate schemes in the third subset do not contain the new BOQ features, performing feature matching on all candidate schemes in the third subset according to the existing network BOQ features to obtain a fourth subset, calculating the feature matching degree of each candidate scheme in the fourth subset, and sequencing the candidate schemes in the fourth subset according to the feature matching degrees to obtain a candidate scheme subset. The reverse subtraction can avoid the problem of combined explosion caused by excessive scenes in forward evolution logic, and obviously reduces the number of obtained candidate scheme subsets, thereby obviously improving the speed of designing the scheme of the base station to be designed.

Alternatively, in some embodiments of the present application,

the plan recommending module 805 is specifically configured to calculate, according to the plan evolution path model, an evolution optimization value of each candidate plan in the candidate plan subset through an evolution path optimization algorithm;

the scheme recommending module 805 is further configured to obtain a recommendation degree of each candidate scheme in the candidate scheme subset according to the evolution priority value and the feature matching degree of the candidate scheme corresponding to the evolution priority value;

the scheme recommending module 805 is further configured to rank all candidate schemes in the candidate scheme subset according to the recommendation degree, so as to obtain a recommendation result.

In this embodiment of the application, on the basis of a known scheme evolution path model, after the scheme recommendation module 805 performs processing by using an evolution path optimization algorithm, an evolution optimization value of each candidate scheme in the candidate scheme subset may be obtained, and after the evolution optimization value of the candidate scheme is calculated, the feature matching degree of the candidate scheme is also known, and the scheme recommendation module 805 performs normalization processing on the evolution optimization value and the feature matching degree to obtain a recommendation degree, and sorts all candidate schemes in the candidate scheme subset in a descending manner according to the recommendation degree of each candidate scheme to obtain a recommendation result.

Alternatively, in some embodiments of the present application,

the self-perfecting module 806 is further configured to determine that the candidate solution in the recommendation result belongs to the second subset or the fourth subset;

the self-perfecting module 806 is further configured to determine that all candidate solutions in the recommendation result are perfect candidate solutions when the candidate solutions in the recommendation result belong to the second subset;

the self-perfecting module 806 is further configured to determine, when the candidate solution in the recommendation result belongs to the fourth subset, a device feature of the base station to be designed according to the feature to be used;

the self-perfecting module 806 is further configured to determine whether the candidate solutions in the recommendation result are completely matched with the device characteristics of the base station to be designed one by one;

a self-perfecting module 806, configured to determine that the first candidate is an imperfect candidate when the first candidate does not completely match with the device characteristics of the base station to be designed;

the self-perfecting module 806 is further configured to determine that the second candidate is a perfect candidate when the second candidate is completely matched with the device characteristics of the base station to be designed.

In the embodiment of the application, since the recommendation result may only be the second subset or the fourth subset, if the recommendation result corresponds to the second subset, it indicates that all candidate schemes in the recommendation result are obtained under the condition that hardware is not replaceable, and then it indicates that the device types or the number of all candidate schemes in the recommendation result are consistent with the device types or data of capacity expansion and transformation of the base station to be designed, and all candidate schemes in the recommendation result are perfect candidate schemes; if the candidate solution in the recommendation result belongs to the fourth subset, the self-perfecting module 805 determines the device features of the base station to be designed according to the features to be used, and judges one by one whether the candidate solution in the recommendation result is completely matched with the device features of the base station to be designed, if the first candidate solution is not completely matched with the device features of the base station to be designed, the self-perfecting module 805 determines that the first candidate solution is an incomplete candidate solution, and if the second candidate solution is completely matched with the device features of the base station to be designed, the self-perfecting module 805 determines that the second candidate solution is a perfect candidate solution. The perfect candidate solutions are directly recommended to the client according to the recommendation results without modification, but if the perfect candidate solutions are imperfect candidate solutions, the perfect candidate solutions cannot be directly recommended to the client, so that after the recommendation results are obtained, whether imperfect candidate solutions exist in the recommendation results or not needs to be judged.

Alternatively, in some embodiments of the present application,

the self-perfecting module 805 is further configured to determine a slot of a port to be adjusted in an incomplete candidate scheme;

the self-improvement module 805 is further configured to generate a frequent connection set through a connection rule mining algorithm according to the device connection knowledge graph;

the self-perfecting module 805 is further configured to determine setting information of a slot position of the port to be adjusted according to the frequent connection set;

the self-perfecting module 805 is further configured to adjust a slot of a port to be adjusted in the incomplete candidate scheme according to the setting information, so as to obtain a self-perfecting candidate scheme;

the self-perfecting module 805 is further configured to obtain a base station solution set to be designed according to the self-perfecting candidate solution and the perfecting candidate solution.

In the embodiment of the present application, after the incomplete candidate solution is determined by the self-perfecting module 805, the slot position of the port to be adjusted in the incomplete candidate solution is determined, and under the condition that the device connection knowledge map is known, a frequent connection set is generated by a connection rule mining algorithm, so that the frequent connection set of the slot position of the port to be adjusted is determined, and then a probabilistic graph model (mainly using bayesian network) is adopted, the matched port slot position is used as condition information, and setting information of the port slot to be adjusted is recommended in the frequent connection set of the port slot position to be adjusted, so that the base station schemes to be designed in the base station scheme set to be designed provided to the customer are perfect, and the base station schemes to be designed and improved are not needed, and the requirements of the customer are met better.

Alternatively, in some embodiments of the present application,

the feature extraction module 801 is specifically configured to perform data cleaning processing and data mapping processing on existing network base station information and base station information to be designed;

the feature extraction module 801 is further configured to convert the processed information of the existing network base station and the information of the base station to be designed into a feature to be used.

In the embodiment of the application, the feature extraction module 801 performs data cleaning processing on the existing network base station information and the base station information to be designed, the data cleaning processing mainly includes selecting valid information, removing useless fields (such as comments and the like), and then performing data mapping processing, for example, the device type and name and the like can be analyzed from a device identification number (ID), the existing network base station information and the base station information to be designed perform data cleaning processing and data mapping processing, and the feature extraction module 801 extracts features to be used by changing the file format.

Optionally, as shown in fig. 9, in some embodiments of the present application, the apparatus further includes:

a knowledge base storage module 901, configured to store the historical solution set, the solution evolution path model, the device connection knowledge graph, and the solution set of the base station to be designed in the knowledge base.

In this embodiment of the application, in order to enhance the performance of the base station capacity expansion and transformation scheme design apparatus, the data size of the historical scheme library needs to be enhanced, so as to obtain a more complete scheme evolution path model and device connection knowledge graph, and then after obtaining the historical scheme set, the scheme evolution path model, the device connection knowledge graph and the base station scheme set to be designed each time, the knowledge base storage module 901 may store the obtained scheme evolution path model, device connection knowledge graph and device connection knowledge graph into the knowledge base, where the knowledge base is a database or a cloud storage, and the like.

In the above embodiments, a method and an apparatus for designing a base station capacity expansion and transformation scheme are introduced, but the base station capacity expansion and transformation scheme design apparatus of the present application may be operated on a device using a server as an entity apparatus, which is specifically as follows:

referring to fig. 10, an embodiment of the present application provides a server, including:

a processor 1001, a transceiver 1002, and a memory 1003, wherein the memory 1003 may be used to store codes executed by the processor 1001;

the processor 1001, the transceiver 1002, and the memory 1003 are connected by a bus system 1004;

the transceiver 1002 is configured to acquire information of a current network base station, information of a base station to be designed, and a history scheme set;

the processor 1001 is further configured to perform data processing on the information of the current network base station and the information of the base station to be designed, so as to obtain a feature to be used;

the processor 1001 is further configured to perform algorithm training learning on the historical scheme set to obtain a scheme evolution path model and a device connection knowledge graph;

the processor 1001 is further configured to select a candidate scheme subset from the historical scheme set through a reverse subtraction algorithm according to the feature to be used;

the processor 1001 is further configured to perform recommendation degree ranking processing on all candidate schemes in the candidate scheme subset according to the scheme evolution path model to obtain a recommendation result;

the processor 1001 is further configured to, when an incomplete candidate exists in the recommendation result, perform self-improvement processing on the incomplete candidate according to the device connection knowledge graph to obtain a to-be-designed base station solution set.

In the embodiment of the application, when a base station needs to be subjected to capacity expansion transformation, a base station design delivery person inputs current network base station information and information of a base station to be designed into a base station capacity expansion transformation scheme design device, a transceiver 1002 can acquire the current network base station information and the information of the base station to be designed which are input by the base station design delivery person, a processor 1001 performs data processing on the current network base station information and the information of the base station to be designed to acquire a feature to be used, acquires a historical scheme set from a network or a database of the device through the input of the base station design delivery person, the processor 1001 performs algorithm training and learning on the historical scheme set to acquire a scheme evolution path model and a device connection knowledge map, and performs feature matching on the feature to be used on the historical scheme set through a reverse subtraction algorithm according to the feature to be used so as to select a candidate scheme subset, the processor 1001 may deduce an evolution path of the capacity expansion transformation scheme of the base station to be designed based on the scheme evolution path model, thereby calculating a recommendation degree of each candidate scheme in the candidate scheme subset, rank the candidate schemes in the candidate scheme subset according to the obtained recommendation degree, obtain a recommendation result, when an incomplete candidate scheme exists in the recommendation result, the processor 1001 may know a connection rule between devices according to a device connection knowledge map, implement self-perfection processing on the incomplete candidate scheme according to the connection rule between devices, enable the incomplete candidate scheme to completely match the base station to be designed, merge the incomplete candidate scheme after self-perfection processing and the originally complete candidate scheme into a set, and obtain a base station to be designed scheme set. By using a reverse subtraction algorithm to obtain the candidate scheme subset, the problem of combined explosion caused by excessive scenes in forward evolution logic can be avoided, and the number of candidate schemes is obviously reduced; performing recommendation degree sorting processing on all candidate schemes in the candidate scheme subset according to the scheme evolution path model to obtain a recommendation result, so that the candidate schemes can be obtained more accurately and quickly; and performing self-perfection processing on the imperfect candidate scheme according to the device connection knowledge graph, so that the candidate scheme which better meets the requirements of customers can be provided. Therefore, the speed, the efficiency and the accuracy of the scheme design of the base station to be designed are improved.

The present application further provides a computer-readable storage medium, where instructions are stored in the computer-readable storage medium, and when the instructions are run on a computer, the computer is enabled to execute the method for designing the base station capacity expansion modification scheme described in the foregoing embodiment.

The present application further provides a computer program product containing instructions, which when run on a computer, causes the computer to execute the method for designing the base station capacity expansion modification scheme described in the above embodiments.

In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product.

The computer program product includes one or more computer instructions. When loaded and executed on a computer, cause the processes or functions described in accordance with the embodiments of the application to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website site, computer, server, or data center to another website site, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can store or a data storage device, such as a server, a data center, etc., that is integrated with one or more available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.

It should be understood that, in the various embodiments of the present application, the sequence numbers of the above-mentioned processes do not mean the execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

The above embodiments are only used for illustrating the technical solutions of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for designing a base station capacity expansion transformation scheme is characterized by comprising the following steps:

when an incomplete candidate scheme exists in the recommendation result, performing self-improvement processing on the incomplete candidate scheme according to the device connection knowledge map to obtain a base station scheme set to be designed;

selecting a candidate scheme subset from the historical scheme set through a reverse subtraction algorithm according to the features to be used, wherein the candidate scheme subset comprises:

obtaining a Radio Frequency (RF) characteristic, a new project amount list BOQ characteristic, a current network BOQ characteristic and a hardware replacement condition characteristic according to the characteristic to be used;

judging whether the hardware replacement condition characteristic is hardware replaceable or not according to the hardware replacement condition characteristic;

and calculating the feature matching degree of each candidate scheme in the second subset, and sequencing the schemes in the second subset according to the feature matching degree to obtain a candidate scheme subset.

2. The method of claim 1, wherein the performing algorithm training learning on the historical solution set to obtain an evolving knowledge-graph and a device connection knowledge-graph comprises:

3. The method of claim 2, wherein selecting a subset of candidate solutions from the set of historical solutions by a reverse subtraction algorithm according to the feature to be used comprises:

4. The method according to claim 3, wherein the performing recommendation ranking processing on all candidate solutions in the candidate solution subset according to the solution evolution path model to obtain a recommendation result comprises:

obtaining the recommendation degree of each candidate scheme in the candidate scheme subset according to the evolution preference value and the feature matching degree of the candidate scheme corresponding to the evolution preference value;

5. The method of claim 4, further comprising:

6. The method of claim 5, wherein the performing self-perfection processing on the imperfect candidate solutions according to the device connection knowledge graph to obtain a solution set of base stations to be designed comprises:

7. The method according to any one of claims 1 to 6, wherein the performing data processing on the information of the existing network base station and the information of the base station to be designed to obtain the feature to be used comprises:

8. The method of claim 7, further comprising:

9. The utility model provides a base station dilatation transformation scheme design device which characterized in that includes:

the self-perfecting module is used for carrying out self-perfecting processing on the imperfect candidate scheme according to the device connection knowledge graph to obtain a base station scheme set to be designed when the imperfect candidate scheme exists in the recommendation result;

the scheme screening module is specifically used for obtaining a Radio Frequency (RF) characteristic, a new project amount list BOQ characteristic, a current network BOQ characteristic and a hardware replacement condition characteristic according to the to-be-used characteristic;

the scheme screening module is further configured to determine whether the hardware replacement condition characteristic is hardware replaceable according to the hardware replacement condition characteristic;

the scheme screening module is further configured to calculate a feature matching degree of each candidate scheme in the second subset, and rank the schemes in the second subset according to the feature matching degree to obtain a candidate scheme subset.

10. The apparatus of claim 9,

11. The apparatus of claim 10,

12. The apparatus of claim 11,

the scheme recommending module is further configured to obtain a recommendation degree of each candidate scheme in the candidate scheme subset according to the evolution preference value and the feature matching degree of the candidate scheme corresponding to the evolution preference value;

13. The apparatus of claim 12,

14. The apparatus of claim 13,

15. The apparatus according to any one of claims 9 to 14,

16. The apparatus of claim 15, further comprising:

17. A server, comprising:

the processor is used for performing self-perfection processing on the imperfect candidate scheme according to the device connection knowledge graph to obtain a base station scheme set to be designed when the imperfect candidate scheme exists in the recommendation result;

the processor is specifically used for obtaining a Radio Frequency (RF) characteristic, a new project amount list BOQ characteristic, a current network BOQ characteristic and a hardware replacement condition characteristic according to the to-be-used characteristic;

the processor is further configured to perform feature matching on all candidate solutions in the historical solution set according to the RF features and the new BOQ features to obtain a first subset;

the processor is further configured to determine whether the hardware replacement condition characteristic is hardware replaceable according to the hardware replacement condition characteristic;

the processor is further configured to, when the hardware replacement condition feature is not hardware replaceable, perform feature matching on all candidate solutions in the first subset according to the current network BOQ feature to obtain a second subset;

the processor is further configured to calculate a feature matching degree of each candidate scheme in the second subset, and rank the schemes in the second subset according to the feature matching degree to obtain a candidate scheme subset.

18. A computer-readable storage medium comprising instructions that, when executed on a computer, cause the computer to perform the base station capacity expansion modification scheme designing method of any one of claims 1 to 8.