CN116260730A - Geographic information service evolution particle swarm optimization method in multi-edge computing node - Google Patents
Geographic information service evolution particle swarm optimization method in multi-edge computing node Download PDFInfo
- Publication number
- CN116260730A CN116260730A CN202310543371.3A CN202310543371A CN116260730A CN 116260730 A CN116260730 A CN 116260730A CN 202310543371 A CN202310543371 A CN 202310543371A CN 116260730 A CN116260730 A CN 116260730A
- Authority
- CN
- China
- Prior art keywords
- service
- edge computing
- particle
- computing node
- chromosome
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 239000002245 particle Substances 0.000 title claims abstract description 155
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000005457 optimization Methods 0.000 title claims abstract description 21
- 210000000349 chromosome Anatomy 0.000 claims abstract description 117
- 230000006978 adaptation Effects 0.000 claims abstract description 40
- 238000004364 calculation method Methods 0.000 claims description 33
- 238000012937 correction Methods 0.000 claims description 17
- 238000012545 processing Methods 0.000 claims description 16
- 230000008569 process Effects 0.000 claims description 13
- 108090000623 proteins and genes Proteins 0.000 claims description 11
- 230000035772 mutation Effects 0.000 claims description 8
- 230000009191 jumping Effects 0.000 claims description 3
- 238000004891 communication Methods 0.000 description 17
- 238000010586 diagram Methods 0.000 description 6
- 238000011156 evaluation Methods 0.000 description 5
- 230000005540 biological transmission Effects 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000012804 iterative process Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/142—Network analysis or design using statistical or mathematical methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- Algebra (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Probability & Statistics with Applications (AREA)
- Pure & Applied Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a geographic information service evolution particle swarm optimization method in a multi-edge computing node, which comprises the steps of establishing an evolution particle swarm algorithm frame aiming at geographic information edge service, carrying out particle chromosome coding of an evolution particle swarm algorithm under limited edge computing nodes, comprehensively considering service history quality log records and edge computing node performances, correcting service quality of concrete realization service, balancing edge computing node distribution conditions in concrete realization service chain combination, obtaining comprehensive adaptation values of each concrete realization service chain combination, updating particle chromosomes, and outputting the global optimal position of a current particle chromosome swarm as an optimal scheme of each service chain combination.
Description
Technical Field
The invention belongs to the field of geographic information science, and relates to a geographic information service evolution particle swarm optimization method in a polygonal computing node.
Background
Disruption of the public communication network by natural disasters presents a significant challenge to disaster data observation, transmission, processing, and application during emergency response. In emergency communication networks, the performance and stability of edge computing nodes is often weak due to limited communication and computing resources, directly affecting the quality of service of geographic information services based on emergency communication and emergency monitoring of the computing environment. The research on constructing a high-efficiency and reliable geographic information edge service combination method in a disaster emergency-oriented communication network (such as an ad hoc communication network) constructed in a limited area is a problem which is urgently solved by the current demand, and is a key bottleneck for supporting the breakthrough of extreme disaster emergency needs. How to solve the problems is the technical problem to be solved by the invention.
The Web service combination optimization method mainly comprises an exhaustion method, a linear programming method, an intelligent optimization method, a machine learning method and the like. The exhaustion algorithm compares the merits of all possible solutions by directly searching the solution space, and finally obtains the optimal solution. The linear programming method adjusts the service quality constraint parameters through linear weighting, and solves the optimization scheme. The intelligent optimization algorithm is a generic name for a series of algorithms such as genetic algorithm, ant colony algorithm, particle swarm algorithm and the like which solve the optimal combination by simulating the development rule of the nature or the physical world. The machine learning method is mainly based on deep reinforcement learning, and the quality of service evaluation is predicted by training a recurrent neural network model. The intelligent optimization algorithm is widely applied because of simple realization and small calculation amount, and a feasible solution meeting the service quality constraint can be found in a limited time. However, the above methods are all researches on a Web service combination method in a traditional cloud computing mode, and candidate services are huge in scale and almost do not consider network and performance limitations. In the edge environment, the communication capability is limited, and the performance and stability of the edge computing node are generally weak, so that the method cannot perform good geographic information service combination in the edge environment.
Disclosure of Invention
According to the defects of the prior art, the invention aims to provide the geographic information service evolution particle swarm optimization method in the multi-edge computing node, so that an optimal scheme of geographic information edge service chain combination can be obtained in a scene with limited communication and poor performance and stability of the edge computing node in an edge environment, and the execution efficiency, stability and reliability of the geographic information service chain of the multi-edge computing node in the edge computing environment are improved.
In order to solve the technical problems, the invention adopts the following technical scheme:
the geographic information service evolution particle swarm optimization method in the multi-edge computing node comprises the following steps:
step S1, establishing an evolutionary particle swarm algorithm framework aiming at geographic information edge service;
step S2, under the condition of limited edge computing nodes, particle chromosome coding of an evolutionary particle swarm algorithm is carried out, each particle chromosome is set to represent a concrete implementation service chain combination, each gene position on the particle chromosome represents a concrete implementation service, the value range of each gene position is a candidate edge computing node set of the concrete implementation service, and each particle chromosome codes the concrete implementation of the concrete implementation service chain combination represented by the particle chromosome in the actual edge computing node combination;
s3, comprehensively considering service history quality log records and edge computing node performances, correcting service quality of the concrete implementation service, balancing edge computing node distribution conditions in the concrete implementation service chain combination, and obtaining comprehensive adaptation values of each concrete implementation service chain combination;
s4, updating the particle chromosome;
and S5, if the particle chromosomes meet the optimized termination condition, outputting an optimal scheme of each particle chromosome, and if the optimal termination condition is not met, performing evolutionary operation on the particle chromosomes, and then jumping to the step S3 until the optimal termination condition is met.
Further, in step S1, a set of edge computing nodes, a service chain combination, an implementation-specific service chain combination, and a particle chromosome population are set.
Further, a set k= {1,2, …, K, …, n } of edge computing nodes is defined, K is the number of edge computing nodes, and n is the number of edge computing nodes;
defining a service chain combination sc= { S 1 ,S 2 ,…, S j ,…,S m}, wherein ,Sj For the j-th abstract service, m is the number of abstract services;
definition of concrete implementation services s j,k Representing abstract services S j A concrete implementation on the edge computing node k;
defining the concrete implementation service chain combination sc= { s 1,k1 , s 2,k2 , …, s j,kj , …, s m,km , k 1 , k 2 , …, k j , …, k m E K }, where s j,kj For abstract services S j Calculating node k at edge j In the specific implementation, sc is set as the particle chromosome;
definition of particle chromosome population Swarm t ={sc 1 t , sc 2 t , …, sc q t , …, sc M t Wherein t is the iteration number of the particle chromosome population, q is the particle chromosome number, M is the particle chromosome population size, sc q t The results of the t-th particle chromosome population iteration for the q-th particle chromosome.
Further, in step S3, historical execution data of each specific implementation service on each edge computing node is counted, and each edge computing node is counted sequentially until all edge computing nodes are counted, until all specific implementation services are counted, and service quality of each specific implementation service is output.
Further, in step S3, the quality of service of the single implementation service is corrected according to the performance of the edge computing node, and then:
wherein ,is specifically toImplementing services s j,k I-th quality of service index of +.>Implementation service s after correcting performance of edge computing node j,k I-th quality of service index of +.>To embody the service s j,k The edge at which is computed the computational performance weight of node k.
Further, in step S3, the distribution situation of the edge computing nodes in the service chain combination process is balanced by the edge computing node balancing operator;
specifically, the calculation formula of the edge calculation node equalization weight BW (k):
wherein counts nodes (k) Is the number of occurrences of an edge compute node k in a concrete implementation service chain combination,for realizing the service s after the performance correction of the edge computing node and the equalization processing of the edge computing node j,k Is the i-th quality of service index of NW (k) for the implementation specific service s j,k The computing performance weight of the node k is calculated at the edge, and BW (k) is the equalizing weight of the node k is calculated at the edge.
Further, in step S3, the i-th quality of service index embodying the service chain combination is expressed as:
wherein ,to embody the service s j,k I-th quality of service index of +.>For the calculation formula of the quality of service index i logically related to the business processes of the service chain combination +.>An ith quality of service indicator for a specific implementation of service chain combination sc;
the calculation formula of the single service quality index of the concrete realization service chain combination after the edge calculation node calculation performance correction and the edge calculation node equalization is shown as a formula (5):
wherein ,for realizing the service s after the performance correction of the edge computing node and the equalization processing of the edge computing node j,k I-th quality of service index of +.>And (3) the ith service quality index of the service chain combination sc is specifically realized after the performance correction of the edge computing node and the equalization processing of the edge computing node.
Further, in step S3, comprehensive adaptation value calculation is performed by integrating various service quality indexes of the service chain:
wherein ,6 QoS index values for implementing service chain combination respectively>Is a weight vector corresponding to each QoS index value, and +.>F (sc) is a comprehensive adaptation value for realizing the service chain combination sc;
after the performance correction of the edge computing nodes and the equalization treatment of the edge computing nodes, the comprehensive adaptation value calculation is carried out on all service quality indexes of the service chain by comprehensive specific implementation:
wherein ,6 service quality index values of a concrete realization service chain combination sc after edge computing node performance correction and edge computing node equalization treatment are respectively +.>The comprehensive adaptation value of the service chain combination sc is specifically realized after the performance correction of the edge computing nodes and the equalization processing of the edge computing nodes.
Further, in step S4, the particle chromosome historic optimal position and the particle chromosome historic optimal comprehensive adaptation value are updated, and the particle chromosome group global optimal position and the particle chromosome group global optimal comprehensive adaptation value are updated.
Further, in step S5, the evolution operation includes a crossover operation, a mutation operation, and an evolution reversal operation.
Compared with the prior art, the invention has the following advantages and beneficial effects:
according to the geographic information service evolution particle swarm optimization method in the multi-edge computing nodes, service history quality log records and edge computing node performances are comprehensively considered based on the edge computing node performances, service quality of specific realization services is corrected, distribution conditions of the edge computing nodes in service chain combinations are balanced, comprehensive adaptation values of each specific realization service chain combination are obtained, and service chain combinations with better quality and efficiency are constructed.
According to the geographic information service evolution particle swarm optimization method in the multi-edge computing nodes, based on the particle swarm algorithm, the service quality of the specific implementation service is corrected, the distribution condition of the edge computing nodes in the service chain combination is balanced, particle chromosomes are updated, the optimal scheme of each particle chromosome is output, the optimal specific implementation service chain combination is obtained, the method can be suitable for scenes with limited communication and poor performance and stability of the edge computing nodes in the edge environment, and the service quality of the specific implementation of abstract service in the edge environment on the actual edge computing nodes is greatly improved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application. The exemplary embodiments of the present invention and the descriptions thereof are for explaining the present invention and do not constitute an undue limitation of the present invention. In the drawings:
FIG. 1 is a flow chart of a method for optimizing a geographic information service evolution particle swarm in a multi-edge computing node of the present invention;
FIG. 2 is a schematic diagram of a geographic information edge service flow in accordance with one embodiment of the present invention;
FIG. 3 is a graphical representation of a combined particle chromosome of a geographic information service in accordance with one embodiment of the invention;
FIG. 4 is a schematic diagram of a crossover operation in one embodiment of the invention;
FIG. 5 is a schematic diagram of an evolutionary reversal operation in one embodiment of the invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the present application will be described in further detail with reference to the accompanying drawings, and the described embodiments should not be construed as limiting the present application, and all other embodiments obtained by those skilled in the art without making any inventive effort are within the scope of the present application.
In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is to be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments and can be combined with one another without conflict.
In the following description, the terms "first", "second", "third" and the like are merely used to distinguish similar objects and do not represent a particular ordering of the objects, it being understood that the "first", "second", "third" may be interchanged with a particular order or sequence, as permitted, to enable embodiments of the application described herein to be practiced otherwise than as illustrated or described herein.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of the present application only and is not intended to be limiting of the present application.
Before further describing embodiments of the present application in detail, the terms and expressions that are referred to in the embodiments of the present application are described, and are suitable for the following explanation.
1. Geographic information edge service: or edge geographic information services, in particular geographic information services deployed on edge computing nodes.
2. Service chain combination: the single geographic information edge service is organized according to a certain workflow to fulfill more complex task requirements.
3. Edge computing node: the system refers to a computer or a server with certain computing capacity and communication capacity and capable of providing geographic information edge service to the outside.
4. Abstract services: to accomplish a task (e.g., dam scene composite perception) there is a specific workflow model that is composed of a plurality of sub-services. In the workflow model, each sub-service is only descriptive information of service features such as functions, inputs, outputs and the like of the service, and has virtually no service capability, so that the sub-service is called an abstract service.
5. The specific realization service: representing the concrete implementation of the abstract service on the actual edge computing node.
6. The service chain combination is realized: representing a combination of concrete implementations of a plurality of concrete implementation services on a plurality of actual edge computing nodes.
FIG. 2 is a schematic diagram of a geographic information edge service flow in accordance with one embodiment of the present invention. The oval frame represents an edge environment under the same communication network, wherein a limited number of edge computing nodes capable of providing geographic information services exist, each edge computing node is used for realizing a plurality of concrete realization services, the concrete realization services can interact through the network, and the concrete realization services are processing services or data services; in order to realize a specific task, a plurality of processing services and data services are organized according to a certain flow, and meanwhile, in order to avoid excessive crash of a certain edge computing node, each geographic information edge service is dispersed into the whole edge environment to operate, solid arrows in the figure represent the sequence of the processing services, and dotted arrows represent the flow direction of the data services.
In an edge computing environment, particularly an emergency communication environment, the combination of service chains in a geographic information edge service needs to consider the following influencing factors: the communication capability between the edge computing nodes is limited, and the computing capability of the edge computing nodes and the service quality of the geographic information service are limited.
The optimization of the geographic information service chain combination in the edge environment in the related technology is research on a Web service combination method in a traditional cloud computing mode, and the candidate service is huge in scale and almost does not consider network and performance limitations. However, in the edge environment, the communication capability is limited, and the performance and stability of the edge computing node are generally weak, so that the quality of the geographic information service chain combination in the edge environment cannot be truly improved by the related technology.
In view of the above technical problems, the embodiments of the present application provide a method for optimizing a geographic information service evolution particle swarm in a multi-edge computing node, which can be applied to a scene with limited communication and poor performance and stability of the edge computing node in an edge environment, and calculate an optimal scheme of geographic information edge service combination on the basis of the scene.
The geographic information service evolution particle swarm optimization method in the multi-edge computing node provided by the invention, as shown in figure 1, comprises the following steps:
step S1, establishing an evolutionary particle swarm algorithm framework aiming at geographic information edge service;
step S2, under the condition of limited edge computing nodes, particle chromosome coding of an evolutionary particle swarm algorithm is carried out, each particle chromosome is set to represent a concrete implementation service chain combination, each gene position on the particle chromosome represents a concrete implementation service, the value range of each gene position is a candidate edge computing node set of the concrete implementation service, and each particle chromosome codes the concrete implementation of the concrete implementation service chain combination represented by the particle chromosome in the actual edge computing node combination;
s3, comprehensively considering service history quality log records and edge computing node performances, correcting service quality of the concrete implementation service, balancing edge computing node distribution conditions in the concrete implementation service chain combination, and obtaining comprehensive adaptation values of each concrete implementation service chain combination;
s4, updating the particle chromosome;
and S5, if the particle chromosomes meet the optimized termination condition, outputting an optimal scheme of each particle chromosome, and if the optimal termination condition is not met, performing evolutionary operation on the particle chromosomes, and then jumping to the step S3 until the optimal termination condition is met.
According to the geographic information service evolution particle swarm optimization method in the multi-edge computing node, service quality of the concrete realization service is corrected by comprehensively considering service history quality log records and edge computing node performances based on the edge computing node performances, the distribution situation of the edge computing nodes in the concrete realization service chain combination is balanced, the comprehensive adaptation value of each concrete realization service chain combination is obtained, and the service chain combination with better quality and efficiency is constructed.
According to the geographic information service evolution particle swarm optimization method in the multi-edge computing nodes, based on the particle swarm algorithm, the service quality of the concrete implementation service is corrected, the distribution condition of the edge computing nodes in the concrete implementation service chain combination is balanced, particle chromosomes are updated, an optimal scheme of each particle chromosome is output, the optimal concrete implementation service chain combination is obtained, the method can be suitable for scenes with limited communication and poor performance and stability of the edge computing nodes in the edge environment, and the service quality of concrete implementation of abstract service on the actual edge computing nodes in the edge environment is greatly improved.
In step S1, the present invention sets an edge computing node set, a service chain combination, a specific implementation service chain combination, and a particle chromosome group.
Specifically, a set k= {1,2, …, K, …, n } of edge computing nodes is defined, K is the number of edge computing nodes, and n is the number of edge computing nodes;
defining a service chain combination sc= { S 1 ,S 2 ,…, S j ,…,S m}, wherein ,Sj For the j-th abstract service, m is the number of abstract services;
definition of concrete implementation services s j,k Representing abstract services S j A concrete implementation on the edge computing node k;
defining the concrete implementation service chain combination sc= { s 1,k1 , s 2,k2 , …, s j,kj , …, s m,km , k 1 , k 2 , …, k j , …, k m E K }, where s j,kj For abstract services S j Calculating node k at edge j In the specific implementation, sc is set as the particle chromosome;
definition of particle chromosome population Swarm t ={sc 1 t , sc 2 t , …, sc q t , …, sc M t Wherein t is the iteration number of the particle chromosome population, q is the particle chromosome number, M is the particle chromosome population size, i.e. the total number of particle chromosomes, sc q t The results of the t-th particle chromosome population iteration for the q-th particle chromosome.
The invention sets the maximum iteration number asThe variation probability p_mutation=0.2 is set, the positions and speeds of all particle chromosomes in the particle chromosome group are initialized in a random mode, the historical optimal comprehensive adaptation value of all particle chromosomes is set to be 0, and the global optimal comprehensive adaptation value of the particle chromosome group is set to be 0.
In one embodiment of the present invention, as shown in FIG. 3, each column in the figure represents a specific implementation service, such as a hyperspectral data service, a oblique photogrammetry data service, a mode filtering service, an earth surface debris extraction service, a slope processing service, and a classification and change detection service; in each column, the top most is a concrete implementation service name, and the lower 1,2, …,9 represents a candidate edge computing node set capable of providing a corresponding concrete implementation service; the horizontal gray-bottom dashed border represents one possible implementation of the service chain combination in the actual edge computing node combination.
The actual executable concrete service is a combination of "concrete implementation service name+edge computing node", and then a concrete implementation of one concrete implementation service chain combination illustrated in fig. 3 in the actual edge computing node combination is fully described as: { hyperspectral data services on edge computing node 1, oblique photogrammetry data services on edge computing node 5, mode filtering services on edge computing node 4, surface clutter extraction services on edge computing node 7, slope processing services on edge computing node 3, classification and change detection services on edge computing node 5 }. Meanwhile, as the sequence of the concrete implementation service in the list is unchanged, the integer sequence is used for encoding the edge computing nodes, and the concrete implementation of the concrete implementation service chain combination in the actual edge computing node combination is further simplified to {1,5,4,7,3,3,5}, namely particle chromosome encoding.
In the invention, in step S3, firstly, the service quality of a single geographic information edge service is obtained, and the service quality statistics information of concrete implementation of each abstract service on an actual edge computing node is obtained, namely, the service quality of each concrete implementation service is obtained, the historical execution data of each concrete implementation service on each edge computing node is counted, and the calculation nodes are counted sequentially until all edge computing nodes are counted until all concrete implementation services are counted, and the service quality of each concrete implementation service is output.
Secondly, correcting the service quality of the single concrete implementation service according to the performance of the edge computing node, and then:
wherein ,to embody the service s j,k I-th quality of service index of +.>Implementation service s after correcting performance of edge computing node j,k I-th quality of service index of +.>To embody the service s j,k The edge at which is computed the computational performance weight of node k.
The calculation performance weight of each edge calculation node takes the comprehensive evaluation score as an index, and the comprehensive evaluation score of each edge calculation node is standardized.
In the standardization process, the comprehensive evaluation score is selected as a hardware comprehensive score of a border computing node (server) by a third party software (such as Lu Da engineer, CPU-Z and the like) with the same version of the software, and the original data is subjected to linear transformation so that all the data fall between 0 and 1 (whether the original data are positive values or negative values).
The computing performance of the edge computing nodes may vary significantly due to the edge environment. The strong edge computing nodes are high in performance parameters, the situation that edge computing nodes are aggregated (namely, most of specific implementation services in the specific implementation service chain combination are on the same edge computing node) easily occurs in the optimization process, excessive concurrency of the edge computing nodes with the aggregated services is caused in the optimization process, the quality of all specific implementation services on the edge computing nodes is seriously affected, and finally, the phenomenon that the implementation effect of the actual specific implementation service chain combination is drastically reduced is caused. And balancing the distribution condition of the edge computing nodes in the service chain combination process through an edge computing node balancing operator.
Specifically, the calculation formula of the edge calculation node equalization weight BW (k):
wherein counts nodes (k) Is the number of occurrences of an edge computing node k in a particular implementation service chain combination, e.g., particular implementation service chain combination { s } 1,6 ,s 2,2 ,s 3,6 ,s 4,9 ,s 5,8 ,s 6,3 ,s 7,2 Count corresponding to edge computing node 2 in the sequence nodes (2) =2, count corresponding to edge compute node 3 nodes (3)=1,For realizing the service s after the performance correction of the edge computing node and the equalization processing of the edge computing node j,k Is the i-th quality of service index of NW (k) for the implementation specific service s j,k The computing performance weight of the node k is calculated at the edge, and BW (k) is the equalizing weight of the node k is calculated at the edge.
And particularly, the service chain combination single service quality index aggregation calculation is realized. The service chain combination in actual execution is formed by combining specific implementation services possibly located on different edge computing nodes according to a certain business flow logic.
The ith service quality index of the concrete implementation service chain combination is expressed as:
wherein ,to embody the service s j,k I-th quality of service index of +.>For the calculation formula of the quality of service index i logically related to the business processes of the service chain combination +.>The ith quality of service indicator for the specific implementation of the service chain combination sc.
Correspondingly, a calculation formula of a single service quality index of a specific implementation service chain combination after edge computing node computing performance correction and edge computing node equalization is shown as a formula (5):
wherein ,for the i-th quality of service indicator for a particular implementation service sj, k calculated according to equation (3),and (3) the ith service quality index of the service chain combination sc is specifically realized after the performance correction of the edge computing node and the equalization processing of the edge computing node.
And finally, comprehensively and specifically realizing each service quality index of the service chain to perform comprehensive adaptation value calculation. And optimizing by taking the service quality of the concrete realization service chain combination as a comprehensive adaptation value, wherein a calculation formula of the comprehensive adaptation value is shown in a formula (6).
wherein And respectively calculating 6 service quality index values of the concrete implementation service chain combination through the formula (4). />Is a weight vector corresponding to each QoS index value, and +.>. F (sc) is a comprehensive adaptation value of the implementation-specific service chain combination sc, and the higher the comprehensive adaptation value is, the higher the evaluation of the implementation-specific service chain combination is:
the above method only refers to the historical service quality information of the specific implementation service chain combination, and lacks consideration on the performance of the edge computing node. After the computing capability attribute of the edge computing node is introduced according to the modes of the formulas (1) - (3) of the invention, the comprehensive adaptation value of the service chain combination is specifically realizedThe calculation formula of (2) is as follows:
wherein ,the 6 qos index values calculated by equation (5) are respectively used for the specific implementation of the service chain combination sc.
Wherein, response time (abbreviated as RT): the task calculation process is performed from the time when a user initiates a service request to a service provider, to the time when the server returns the user service execution result, including the data (request) transmission time and the actual calculation task execution time. Reliability (Reliability, abbreviated as Rel): the service callback success probability is defined as the ratio of the number of successful service call times to the total number of times. Availability (availabilities, abbreviated as Ava): defined as the probability that the Web service is accessible. Security (Sec): and comprehensively evaluating the data and information security in the process of Web service data transmission and execution. Cost of execution (Cost): the economic charge paid to perform Web services. Reputation (Rep): user satisfaction with the service is typically measured by a user score.
In the present invention, in step S4, the particle chromosome history optimal position and the particle chromosome history optimal comprehensive adaptation value are updated, and the particle chromosome group global optimal position and the particle chromosome group global optimal comprehensive adaptation value are updated.
Specifically, the current integrated adaptation value of the particle chromosome representing each service chain combination is compared with the optimal integrated adaptation value of the particle chromosome history, if the current integrated adaptation value is superior to the optimal integrated adaptation value of the particle chromosome history, the optimal position of the particle chromosome history and the optimal integrated adaptation value of the particle chromosome history are updated, the global optimal integrated adaptation value of the particle chromosome representing each service chain combination is compared with the global optimal integrated adaptation value of the particle chromosome group, and if the current integrated adaptation value is superior to the global optimal integrated adaptation value of the particle chromosome group, the global optimal position of the updated particle chromosome group and the global optimal integrated adaptation value of the particle chromosome group are updated.
In the invention, the number of the edge computing nodes is limited under the edge environment, so that the number and the scale of the candidate edge computing node sets of the abstract service are limited, and the gene bit value range representing one abstract service in the service chain combination is very narrow. If the particle-chromosome group is updated by using the speed-position updating formula of the standard particle-group algorithm, slight displacement generated by setting the inertia weight w and the learning factors c1 and c2 in the formula is lost due to discrete rounding operation, so that the particle-chromosome group iterates and oscillates. The operations of inversion, crossover and mutation of chromosome evolution in the particle swarm evolution algorithm are naturally suitable for discrete object operation, and no extra discretization loss is generated. Therefore, the present invention replaces the original particle chromosome population updating method of the standard particle population with the crossover operation, the mutation operation and the evolution reverse operation.
In step S4, the particle chromosome is subjected to an evolution operation including a crossover operation, a mutation operation, and an evolution inversion operation.
Specifically, the crossing operation comprises crossing with a historical optimal position of the particle chromosome itself in an iterative process, crossing with a global optimal position of a current particle chromosome group, wherein the front and back crossing operations both adopt single-point crossing operators, only the current position of the particle chromosome is changed when the crossing operation is carried out, and the historical optimal position of the particle chromosome and the global optimal position of the particle chromosome group are kept unchanged;
based on the integer sequence coding mode, the mutation operation is as follows: generating M x M random number sequences ptrs, wherein M is the size of a particle chromosome group, and M is the number of abstract services;
traversing ptrs, letting ptr x For the xth random number in the sequence of random numbers, compare ptr x And the mutation probability p_mutation, when ptr x <p_mutation, entering the next step; otherwise, continuing to traverse ptrs;
para ptr x Performing a mutation operation on the directed particle chromosome position: randomly taking other integers in the value range;
the evolutionary reversal operation is as follows:
generating r 1 ,r 2 Two random numbers, where r 1 <r 2 ;
For r 1 and r2 The sequence between them is reversed;
judging whether the comprehensive adaptation value of the specific implementation service chain combination after the reversion is improved, if so, reserving the sequence after the reversion; otherwise, the sequence before reversal is restored.
Fig. 4 is a schematic diagram of a crossover operation. The top one is an abstract service name involved in the example service chain combination; every behavior under the particle chromosome, the particle chromosome meaning is consistent with the left annotation; each gene bit in the particle corresponds to the name of the uppermost abstract service one by one, and represents the number of the concrete edge computing node for bearing the abstract service.
The particle chromosomes are crossed twice successively: (1) crossing with a historical optimal position of the particle chromosome in the iterative process; (2) crossing with the global optimal position of the current particle chromosome population. The front and back steps of crossing operation process are the same, and single-point crossing operators are adopted.
Illustrated with a first crossover. Firstly, a random integer is taken to divide the particle chromosome into two sections (for example: {9,8,6,9,8,3,2} is divided into a front section = {9,8,6,9} and a rear section = {8,3,2}, and {6,2,6,9,2,7,6} is divided into a front section = {6,2,6,9} and a rear section = {2,7,6 }), the rear section {8,3,2} of the historical optimal position of the particle chromosome itself (the second crossing is the global optimal position of the particle chromosome group) is used for covering the rear section {2,7,6} of the particle chromosome itself, and the combination is carried out with the front section {6,2,6,9} of the particle chromosome itself to obtain the particle chromosome {6,2,6,9,8,3,2} after the first crossing.
FIG. 5 is a schematic diagram of an evolutionary reversal operation. The top one is an abstract service name involved in the example service chain combination; every behavior under the particle chromosome, the particle chromosome meaning is consistent with the left annotation; each gene bit in the particle corresponds to the name of the uppermost abstract service one by one, and represents the number of the concrete edge computing node for bearing the abstract service.
The evolutionary reversal process is as follows: (1) Generating r 1 ,r 2 Two random numbers, where r 1 <r 2 . (2) For r 1 and r2 The sequence in between is reversed. (3) Judging whether the comprehensive adaptation value of the specific implementation service chain combination after the reversion is improved, if so, reserving the sequence after the reversion; otherwise, the sequence before reversal is restored.
Illustrating: first take two random integers r 1 and r2 (r 1 <r 2 ) Taking r 1 and r2 The order of the genes {6,9,2}, which is reversed (2,9,6 after reversal), the integrated fitness of the reversed particle chromosome {6,2,2,9,6,7,6} is calculated according to step S3 in the present invention, and if the integrated fitness increases, the particle chromosome {6,2,2,9,6,7,6} after reversal of part of the genes is output,otherwise, the particle chromosome {6,2,6,9,2,7,6} before inversion is output.
In summary, the optimization method for the geographic information service evolution particle swarm in the multi-edge computing node provided by the invention can obtain the optimal scheme of the geographic information edge service chain combination in the scene of limited communication and poor performance and stability of the edge computing node in the edge environment, and improves the execution efficiency, stability and reliability of the geographic information service chain of the multi-edge computing node in the edge computing environment.
In step S5, whether the iteration number of the particle chromosome group is smaller than the maximum iteration number T is checked, if T is smaller than or equal to T, the particle chromosome group is updated, and the particle chromosome group iterated is continued; otherwise, the particle chromosome population iteration is finished.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, it is intended that the present invention also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.
Claims (10)
1. The geographic information service evolution particle swarm optimization method in the multi-edge computing node is characterized by comprising the following steps of:
step S1, establishing an evolutionary particle swarm algorithm framework aiming at geographic information edge service;
step S2, under the condition of limited edge computing nodes, particle chromosome coding of an evolutionary particle swarm algorithm is carried out, each particle chromosome is set to represent a concrete implementation service chain combination, each gene position on the particle chromosome represents a concrete implementation service, the value range of each gene position is a candidate edge computing node set of the concrete implementation service, and each particle chromosome codes the concrete implementation of the concrete implementation service chain combination represented by the particle chromosome in the actual edge computing node combination;
s3, comprehensively considering service history quality log records and edge computing node performances, correcting service quality of the concrete implementation service, balancing edge computing node distribution conditions in the concrete implementation service chain combination, and obtaining comprehensive adaptation values of each concrete implementation service chain combination;
s4, updating the particle chromosome;
and S5, if the particle chromosomes meet the optimized termination condition, outputting an optimal scheme of each particle chromosome, and if the optimal termination condition is not met, performing evolutionary operation on the particle chromosomes, and then jumping to the step S3 until the optimal termination condition is met.
2. The method for optimizing a geographic information service evolution particle swarm in a multi-edge computing node according to claim 1, wherein:
in step S1, a set of edge computing nodes, a service chain combination, a specific implementation service chain combination, and a particle chromosome population are set.
3. The method for optimizing a geographic information service evolution particle swarm in a multi-edge computing node according to claim 2, wherein:
defining a set K= {1,2, …, K, …, n } of edge computing nodes, wherein K is the number of the edge computing nodes, and n is the number of the edge computing nodes;
defining a service chain combination sc= { S 1 ,S 2 ,…, S j ,…,S m}, wherein ,Sj For the j-th abstract service, m is the number of abstract services;
definition of concrete implementation services s j,k Representing abstract services S j A concrete implementation on the edge computing node k;
defining the concrete implementation service chain combination sc= { s 1,k1 , s 2,k2 , …, s j,kj , …, s m,km , k 1 , k 2 , …, k j , …, k m E K }, where s j,kj For abstract services S j Calculating node k at edge j In the specific implementation, sc is set as the particle chromosome;
definition of particle chromosome population Swarm t ={sc 1 t , sc 2 t , …, sc q t , …, sc M t Wherein t is the iteration number of the particle chromosome population, q is the particle chromosome number, M is the particle chromosome population size, sc q t The results of the t-th particle chromosome population iteration for the q-th particle chromosome.
4. The method for optimizing a geographic information service evolution particle swarm in a multi-edge computing node according to claim 1, wherein:
in step S3, historical execution data of each specific implementation service on each edge computing node is counted, and each edge computing node is counted sequentially until all edge computing nodes are counted, until all specific implementation services are counted, and service quality of each specific implementation service is output.
5. The method for optimizing a geographic information service evolution particle swarm in a multi-edge computing node according to claim 1, wherein:
in step S3, correcting the service quality of the single implementation service according to the performance of the edge computing node, then:
6. The method for optimizing a geographic information service evolution particle swarm in a multi-edge computing node according to claim 1, wherein:
in step S3, the distribution condition of the edge computing nodes in the service chain combination process is balanced through an edge computing node balancing operator;
specifically, the calculation formula of the edge calculation node equalization weight BW (k):
wherein counts nodes (k) Is the number of occurrences of an edge compute node k in a concrete implementation service chain combination,for realizing the service s after the performance correction of the edge computing node and the equalization processing of the edge computing node j,k Is the i-th quality of service index of NW (k) for the implementation specific service s j,k The computing performance weight of the node k is calculated at the edge, and BW (k) is the equalizing weight of the node k is calculated at the edge.
7. The method for optimizing a geographic information service evolution particle swarm in a multi-edge computing node according to claim 1, wherein:
in step S3, the i-th quality of service index embodying the service chain combination is expressed as:
wherein ,to embody the service s j,k I-th quality of service index of +.>For the calculation formula of the quality of service index i logically related to the business processes of the service chain combination +.>An ith quality of service indicator for a specific implementation of service chain combination sc;
the calculation formula of the single service quality index of the concrete realization service chain combination after the edge calculation node calculation performance correction and the edge calculation node equalization is shown as a formula (5):
wherein ,for realizing the service s after the performance correction of the edge computing node and the equalization processing of the edge computing node j,k I-th quality of service index of +.>And (3) the ith service quality index of the service chain combination sc is specifically realized after the performance correction of the edge computing node and the equalization processing of the edge computing node.
8. The method for optimizing a geographic information service evolution particle swarm in a multi-edge computing node according to claim 1, wherein:
in step S3, comprehensive adaptation value calculation is performed for comprehensively and specifically implementing each service quality index of the service chain:
wherein ,6 QoS index values for implementing service chain combination respectively>Is a weight vector corresponding to each QoS index value, and +.>F (sc) is a comprehensive adaptation value for realizing the service chain combination sc;
after the performance correction of the edge computing nodes and the equalization treatment of the edge computing nodes, the comprehensive adaptation value calculation is carried out on all service quality indexes of the service chain by comprehensive specific implementation:
wherein ,6 service quality index values of a concrete realization service chain combination sc after edge computing node performance correction and edge computing node equalization treatment are respectively +.>The comprehensive adaptation value of the service chain combination sc is specifically realized after the performance correction of the edge computing nodes and the equalization processing of the edge computing nodes.
9. The method for optimizing a geographic information service evolution particle swarm in a multi-edge computing node according to claim 1, wherein:
in step S4, the particle chromosome historic optimal position and the particle chromosome historic optimal comprehensive adaptation value are updated, and the particle chromosome group global optimal position and the particle chromosome group global optimal comprehensive adaptation value are updated.
10. The method for optimizing a geographic information service evolution particle swarm in a multi-edge computing node according to claim 1, wherein:
in step S5, the evolution operation includes a crossover operation, a mutation operation, and an evolution reversal operation.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310543371.3A CN116260730B (en) | 2023-05-15 | 2023-05-15 | Geographic information service evolution particle swarm optimization method in multi-edge computing node |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310543371.3A CN116260730B (en) | 2023-05-15 | 2023-05-15 | Geographic information service evolution particle swarm optimization method in multi-edge computing node |
Publications (2)
Publication Number | Publication Date |
---|---|
CN116260730A true CN116260730A (en) | 2023-06-13 |
CN116260730B CN116260730B (en) | 2023-07-21 |
Family
ID=86686480
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310543371.3A Active CN116260730B (en) | 2023-05-15 | 2023-05-15 | Geographic information service evolution particle swarm optimization method in multi-edge computing node |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116260730B (en) |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8018874B1 (en) * | 2009-05-06 | 2011-09-13 | Hrl Laboratories, Llc | Network optimization system implementing distributed particle swarm optimization |
US20120053970A1 (en) * | 2010-08-25 | 2012-03-01 | International Business Machines Corporation | Systems and methods for dynamic composition of business processes |
KR101669366B1 (en) * | 2015-07-03 | 2016-11-09 | 인하대학교 산학협력단 | System and method for distributed service integration and cost minimization usign genetic algorithm |
CN112118312A (en) * | 2020-09-17 | 2020-12-22 | 浙江大学 | Network burst load evacuation method facing edge server |
CN112181655A (en) * | 2020-09-30 | 2021-01-05 | 杭州电子科技大学 | Hybrid genetic algorithm-based calculation unloading method in mobile edge calculation |
CN112969144A (en) * | 2021-02-02 | 2021-06-15 | 武汉大学 | Micro-service pre-deployment method and system for mobile edge calculation |
WO2022007212A1 (en) * | 2020-07-10 | 2022-01-13 | 南京邮电大学 | Software microservice combination optimization method for edge cloud system |
CN114356548A (en) * | 2021-12-07 | 2022-04-15 | 北京邮电大学 | Dynamic expansion and placement method and device for edge computing service |
CN114862072A (en) * | 2022-07-07 | 2022-08-05 | 武汉大学 | Particle group path search algorithm based on dynamic adjustment of space-time weight |
WO2022267791A1 (en) * | 2021-06-22 | 2022-12-29 | 南京邮电大学 | Workflow scheduling method and system based on multi-objective particle swarm algorithm, and storage medium |
-
2023
- 2023-05-15 CN CN202310543371.3A patent/CN116260730B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8018874B1 (en) * | 2009-05-06 | 2011-09-13 | Hrl Laboratories, Llc | Network optimization system implementing distributed particle swarm optimization |
US20120053970A1 (en) * | 2010-08-25 | 2012-03-01 | International Business Machines Corporation | Systems and methods for dynamic composition of business processes |
KR101669366B1 (en) * | 2015-07-03 | 2016-11-09 | 인하대학교 산학협력단 | System and method for distributed service integration and cost minimization usign genetic algorithm |
WO2022007212A1 (en) * | 2020-07-10 | 2022-01-13 | 南京邮电大学 | Software microservice combination optimization method for edge cloud system |
CN112118312A (en) * | 2020-09-17 | 2020-12-22 | 浙江大学 | Network burst load evacuation method facing edge server |
CN112181655A (en) * | 2020-09-30 | 2021-01-05 | 杭州电子科技大学 | Hybrid genetic algorithm-based calculation unloading method in mobile edge calculation |
CN112969144A (en) * | 2021-02-02 | 2021-06-15 | 武汉大学 | Micro-service pre-deployment method and system for mobile edge calculation |
WO2022267791A1 (en) * | 2021-06-22 | 2022-12-29 | 南京邮电大学 | Workflow scheduling method and system based on multi-objective particle swarm algorithm, and storage medium |
CN114356548A (en) * | 2021-12-07 | 2022-04-15 | 北京邮电大学 | Dynamic expansion and placement method and device for edge computing service |
CN114862072A (en) * | 2022-07-07 | 2022-08-05 | 武汉大学 | Particle group path search algorithm based on dynamic adjustment of space-time weight |
Non-Patent Citations (5)
Title |
---|
付佳;李爱萍;段利国;王舒漫;赵菊敏;: "物联网环境下QoS驱动的服务组合能量优化研究", 小型微型计算机***, no. 09 * |
刘莉平;陈志刚;刘爱心;: "基于粒子群算法的Web服务组合研究", 计算机工程, no. 05 * |
李佳;李海波;: "基于遗传算法的资源服务链构建方法", 小型微型计算机***, no. 09 * |
边馥苓;谭喜成;: "适应于分布式虚拟地理环境服务的对等网络模型研究", 武汉大学学报(信息科学版), no. 11 * |
黄伯虎;段振华;: "量子遗传算法在Web服务选择中的应用", 西安电子科技大学学报, no. 01, pages 1 - 3 * |
Also Published As
Publication number | Publication date |
---|---|
CN116260730B (en) | 2023-07-21 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108989098B (en) | Time delay optimization-oriented scientific workflow data layout method in hybrid cloud environment | |
Wang et al. | A novel reputation-aware client selection scheme for federated learning within mobile environments | |
CN112714032B (en) | Wireless network protocol knowledge graph construction analysis method, system, equipment and medium | |
CN111275172B (en) | Feedforward neural network structure searching method based on search space optimization | |
US20220248237A1 (en) | Neural network circuit remote electrical tilt antenna infrastructure management based on probability of actions | |
CN109413710B (en) | Clustering method and device of wireless sensor network based on genetic algorithm optimization | |
CN114638021B (en) | Security evaluation method for lightweight block chain system of Internet of things | |
CN107895038A (en) | A kind of link prediction relation recommends method and device | |
CN113128432B (en) | Machine vision multitask neural network architecture searching method based on evolution calculation | |
CN113516501A (en) | User communication behavior prediction method and device based on graph neural network | |
CN116050540B (en) | Self-adaptive federal edge learning method based on joint bi-dimensional user scheduling | |
CN115277115A (en) | Method and system for solving robust information propagation problem on network | |
Shapira et al. | Unveiling the type of relationship between autonomous systems using deep learning | |
CN113935398B (en) | Network traffic classification method and system based on small sample learning in Internet of things environment | |
CN116260730B (en) | Geographic information service evolution particle swarm optimization method in multi-edge computing node | |
CN117392483A (en) | Album classification model training acceleration method, system and medium based on reinforcement learning | |
CN107155215A (en) | The distribution method and device of a kind of application home service cluster | |
CN109697511B (en) | Data reasoning method and device and computer equipment | |
CN115941506A (en) | Multi-type service resource arrangement method based on strategy network reinforcement learning | |
Delaplace et al. | Two evolutionary methods for learning bayesian network structures | |
Sasirekha et al. | A comparative analysis of fuzzy-based AHP derived MCDM methods to select the apt heterogeneous wireless network | |
CN115866071A (en) | Method and device for determining reporting path of initial equipment attribute information | |
CN111464343B (en) | Maximum-strain greedy expansion community discovery method and system based on average mutual information | |
Mascherini et al. | M-GA: A genetic algorithm to search for the best conditional Gaussian Bayesian network | |
CN116155755B (en) | Link symbol prediction method based on linear optimization closed sub-graph coding |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |