A benchmark test method and system for optimizing collaboration of network platform manufacturing services

Abstract

The present application relates to a kind of benchmark test method and system of network platform manufacturing service cooperation optimization, comprising the following steps:1.resolve network platform manufacturing service cooperation optimization element relationship, construct network platform manufacturing service cooperation optimization complex network model;2.according to actual problem situation, configure network platform manufacturing service cooperation optimization problem model as needed;3.considering test problem demand, import or generate test data as needed;4.considering test algorithm situation, import or configure the algorithm of solving problem as needed;5.solve the problem multiple times, and record the test results of this problem.The present application can effectively simulate and simulate network platform manufacturing service cooperation optimization environment and process, and can be configured and tested according to the actual needs of user to problem, data and algorithm and compare, to solve the problem of lack of unified benchmark test system tool for previous different problems and algorithm.

Description

Technical Field

[0001] This invention belongs to the field of network platform manufacturing service collaboration optimization technology in service-oriented manufacturing systems, and specifically relates to a benchmark testing method and system for network platform manufacturing service collaboration optimization. Background Technology

[0002] In the research and application of collaborative optimization for manufacturing services on network platforms, participants, managers, and researchers all face the challenge of a lack of benchmark testing systems and tools. The key elements of collaborative optimization for manufacturing services on network platforms can be mainly divided into three categories: manufacturing task requirements, manufacturing service resources, and the collaborative environment of the manufacturing service on the network platform. Manufacturing task requirements generally refer to the demands generated during the manufacturing and production process, including the material distribution process in the workshop, various stages of construction projects, and the execution of computer programs. Each task contains various attributes, including but not limited to priority level, time constraints, and completion quality. Manufacturing service resources generally refer to the resources obtained after manufacturing resources or capabilities have been processed through the manufacturing service system's perception access, virtualization, and service encapsulation, including but not limited to workshop materials, machines, project participants, and computer processors. Task requirements and service resources coexist within a specific collaborative environment of the platform's manufacturing services and are constrained and influenced by that environment. Optimization methods are used to optimize one or more objectives, thereby achieving continuous operation.

[0003] Currently, most participants, managers, and researchers in network platform manufacturing service collaboration employ their own research management methods to study the optimization problem, proposing different problem descriptions, using varying test data, and developing diverse algorithms. However, these studies lack cross-sectional comparison. Therefore, this paper proposes a novel approach to achieve unified modeling and problem description for network platform manufacturing service collaboration optimization. This approach generates test data and configures algorithms according to a standardized process, enabling effective comparison of various current studies on network platform manufacturing service collaboration optimization. This is of great significance for research management capabilities in this field and for ensuring efficient and stable production in the manufacturing sector. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a benchmarking method and system for network platform manufacturing service collaboration optimization. Benchmarking typically refers to establishing a standardized model and data set within a specific research field, enabling researchers and users to conduct horizontal comparisons of a class of problems within the same model and data. This invention effectively solves the lack of a unified descriptive model, test data generation, and algorithm configuration methods in the field of network platform manufacturing service collaboration optimization, enabling horizontal comparative evaluation within related research areas. The method includes: analyzing the relationships between elements in network platform manufacturing service collaboration optimization; constructing a complex network model for network platform manufacturing service collaboration optimization; configuring the network platform manufacturing service collaboration optimization problem model as needed; importing or generating test data as needed; importing or configuring algorithms for solving the problem as needed; solving the test problem multiple times; and recording and storing the test results.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A benchmark testing method for optimizing collaborative manufacturing services on a network platform includes the following steps:

[0007] Step 1: Analyze the relationships between the elements of network platform manufacturing service collaboration optimization and construct a complex network model for network platform manufacturing service collaboration optimization. In the process of network platform manufacturing service collaboration optimization, there are three aspects: providers, demanders and managers. From the perspective of complex network model construction, the model elements are divided into manufacturing service resources, manufacturing service demand and the relationship between supply and demand. Based on the model elements, a complex network model for network platform manufacturing service collaboration optimization is constructed.

[0008] Step 2: Based on the actual problem situation, configure the network platform manufacturing service collaboration optimization problem model as needed; conduct problem configuration and performance evaluation analysis, wherein the problem configuration refers to setting various assumptions and constraints for the optimization problem, and the performance evaluation is to evaluate the overall collaboration process based on the collaboration optimization.

[0009] Step 3: Consider the requirements of the test problem and import or generate test data as needed; First, determine whether the requirements of the test problem have real data. If real data is available, import the real data directly for testing; if real data is not available, simulate and generate the test data; under the premise that the various types of data do not conflict with each other, generate test data that meets the requirements of the test problem.

[0010] Step 4: Consider the test algorithm situation and import or configure the test algorithm for solving the test problem as needed; First, determine whether there is a need to test a specific algorithm for the test problem. If there is a need, import the specific algorithm corresponding to the need; if there is no need to test a specific algorithm, configure the algorithm as needed according to the model of the test problem; the algorithm configured as needed is an intelligent optimization algorithm.

[0011] Step 5: Solve the test problem multiple times and record and store the test results. Considering the random uncertainty of the test problem, if the test object is a problem model, use different test data and solving algorithms to obtain an approximate optimal solution; if the test object is an algorithm, use different test data to compare with other algorithms, such as multi-swarm intelligence algorithms and improved evolutionary algorithms, to obtain the best comparison results.

[0012] Furthermore, step 1 specifically includes:

[0013] (1) For the manufacturing task network T Net = <T,sT,E T > where T represents a complex manufacturing task, sT represents a subtask, and E T This indicates the relationships between subtasks. The description of manufacturing task T includes the number of subtasks of task T. Mission arrival time Task deadline Task cost constraints Task quality constraints Task progress status Task completion status Task value and task completion assessment

[0014] The description of subtask sT includes the subtask type. Subtask deadline Subtask cost constraints Subtask quality constraints Subtask process constraints Subtask progress status Subtask completion status Subtask value Subtask completion assessment

[0015] Relationship between subtasks E T for in, Refers to the process relationship between subtasks. Subtask attribution;

[0016] (2) For manufacturing service network SNet = <S,E S >, where S represents manufacturing services, E S This indicates the relationships between services. The description of the manufacturing service S includes the type of service S. Service Costs Service quality Service efficiency Service capacity Service collaboration efficiency Service reliability Service failure status Service load status Service utility

[0017] Service Relationship E S for in, Refers to service similarity, Indicates the intensity of service collaboration. This refers to the service attribution relationship;

[0018] (3) The collaborative relationship between manufacturing services and tasks is represented as follows: Service S i Subtask sT was executed j Based on the collaborative relationship between manufacturing services and tasks Get the output of the service As shown in equation (1), the profit of the service As shown in equation (2), where, Service S i Execute subtasks sT j Cost:

[0019]

[0020]

[0021] Furthermore, step 2 specifically includes:

[0022] (1) For the model configuration of the network platform manufacturing service collaboration optimization problem, determine the information scope, scheduling timing, scheduling scope and optimization objective of the collaboration optimization problem respectively;

[0023] First, the information scope of this collaborative optimization problem is defined, which includes complete information, partial information, and complete response. Complete information means that the entire set of tasks is known to the scheduling system from the beginning. Partial information means that the scheduling system can only process a portion of the tasks that can be foreseen. Complete response means that the scheduling system can only know the information of the task until it arrives.

[0024] Secondly, the scheduling timing for this collaborative optimization problem needs to be clarified. The scheduling timing includes periodic-driven scheduling and event-driven scheduling. The periodic-driven scheduling refers to scheduling once every fixed time period, while the event-driven scheduling refers to scheduling once whenever a task arrives.

[0025] Subsequently, the scheduling scope of the collaborative optimization problem is determined, which includes local scheduling and global scheduling. Local scheduling refers to selecting only a portion of tasks and services for scheduling in a single scheduling process, while global scheduling refers to considering all tasks and services in a single scheduling process and performing a global solution.

[0026] Finally, the scheduling system determines the scheduling objectives, which are divided into single-objective scheduling and multi-objective scheduling based on the number of objectives.

[0027] (2) Performance evaluation of network platform manufacturing service collaboration optimization refers to the evaluation of the overall system operation, which is divided into internal attributes and external attributes for separate evaluation;

[0028] First, the intrinsic attributes are evaluated, considering the service clustering coefficient of the manufacturing service network, the network throughput, and the network degree distribution; among these... Representative Service S i With S j similarity, Representative Service S i With S j complementarity, Representative Service S i With S j Based on the attribution relationship, the network degree distribution is divided into service similarity distribution. Service complementarity distribution and service attribution distribution The specific calculation methods are shown in equations (3)-(5):

[0029]

[0030]

[0031]

[0032] Secondly, external attributes are evaluated, taking into account system stability, load balancing, system robustness, and system reliability.

[0033] Furthermore, step 3 specifically includes:

[0034] (1) Determine whether the collaborative optimization problem has real data and whether test data is needed. If test data is not needed, import the real data and proceed to the next test.

[0035] (2) Generate test data. First, describe the data related to the test problem. According to different types of test data, set the range and distribution of the corresponding test data to ensure that the test data conforms to the description of the problem during the generation process. Generate data for the service network and the task network respectively.

[0036] (3) For the service network, firstly, generate the service network scale, including the number of services and the number of enterprises; for the enterprise scale, if there is a demand, generate the scale of specific enterprises, including the number of enterprise services, the number of enterprise service clusters, and the capacity of enterprise service clusters, and check whether there is a conflict with the cluster distribution. If there is a conflict, repeat the above service network scale generation process until there is no conflict; then generate service clusters, including the number of service clusters and the service cluster capacity; for enterprise service attributes, if there is a demand, generate the service attributes of specific enterprises, including the number of enterprise service functions, enterprise service cost, enterprise service quality, enterprise service reliability, enterprise service efficiency, and enterprise service collaboration efficiency, and check whether they conflict with the service attribute distribution. If there is a conflict, repeat the above service attribute generation process until there is no conflict; then generate the remaining service attributes that meet the service requirements, including the number of service functions, service cost, service quality, service reliability, service efficiency, and service collaboration efficiency; finally, check whether there are any abnormal services and make modifications.

[0037] (4) For the task network, the first step is to determine whether there are requirements for enterprise task load. If there are requirements, generate enterprise task load, including enterprise task arrival, number of enterprise subtasks, and enterprise subtask size, and check whether they conflict with the task load distribution. If so, repeat the above task load generation process until there is no conflict. Then generate the remaining task load, including task arrival, number of subtasks, and subtask size. For enterprise task attributes, if there are requirements, generate task attributes for specific enterprises, including enterprise task chain length, enterprise task waiting tolerance, enterprise task delay tolerance, enterprise task quality tolerance, and enterprise task price tolerance, and check whether they conflict with the task attribute distribution. If there is a conflict, repeat the above task attribute generation process until there is no conflict. Then generate the remaining task attributes that meet the task requirements, including enterprise task chain length, task waiting tolerance, task delay tolerance, task quality tolerance, and task price tolerance. Finally, check whether there are any abnormal tasks and make modifications.

[0038] Furthermore, step 4 includes:

[0039] (1) Determine whether the collaborative optimization problem requires testing a specific external algorithm or configuring an existing algorithm. If no existing algorithm needs to be configured, import the specific external algorithm and proceed to the next test.

[0040] (2) For the configuration of the intelligent optimization algorithm, the first step is population generation; the population generation includes individual encoding to form a group encoding sequence; each individual in the group encoding sequence contains all the tasks that need to be scheduled, and the encoding sequence of each task contains the schemes of all subtasks, including the subtask start time and subtask scheduling service; then the size of the population is determined.

[0041] (3) The population update mechanism adopts swarm intelligence optimization algorithms, including bee colony algorithm, particle swarm algorithm, and evolutionary algorithm; users adjust the number of iterations and cross-evolution factors of the optimization algorithm to obtain personalized algorithms for testing;

[0042] (4) Set the constraints and termination conditions of the intelligent optimization algorithm. The user sets the range in which the intelligent optimization algorithm searches, the number of search iterations, and the final solution accuracy.

[0043] Furthermore, step 5 specifically includes:

[0044] Based on the manufacturing service collaboration optimization problem model to be tested, the test data, and the test algorithm, the collaboration optimization problem is solved. Users test its performance under different datasets and different algorithms according to different problem models, so as to make an intuitive comparison of the network platform manufacturing service collaboration optimization problem and solution under different backgrounds. After the solution is completed, the historical records are stored and retained, recording the most suitable test data and the optimal intelligent optimization algorithm configuration.

[0045] This invention provides a benchmark testing system for network platform manufacturing service collaboration optimization, comprising: a manufacturing service collaboration benchmark testing system model library, a database, and an algorithm library. The system configures a model based on the manufacturing service collaboration optimization problem to be tested, stores the model in the model library, then retrieves test data adapted to the model from the database, and calls test algorithms adapted to the model from the algorithm library to solve the collaboration optimization problem. The results are then compared and displayed on the benchmark testing system. Simultaneously, the algorithm library stores the test comparison results in the database, and the database provides feedback and corrections to the model based on the quality of the test results.

[0046] The advantages of this invention compared to the prior art are:

[0047] (1) This invention constructs a generalized model for collaborative optimization of network platform manufacturing services based on complex networks. It fully considers the usage needs of participants, managers and researchers in the collaborative optimization process of network platform manufacturing services, analyzes the three types of elements of complex network for collaborative optimization of network platform manufacturing services, and constructs a model for it from three aspects: manufacturing tasks, manufacturing services and supply and demand relationship.

[0048] (2) This invention proposes a problem model construction process for network platform manufacturing service collaboration optimization, specifically including determining its information scope, scheduling timing, scheduling range, and optimization objective. Constructing a network platform manufacturing service collaboration optimization problem based on the above process can effectively simulate real-world problems.

[0049] (3) This invention proposes a set of network platform manufacturing service collaborative optimization test data generation process, which specifically includes analyzing the necessity of test data, defining the data range and distribution, and generating and verifying data as needed, so that users and researchers can generate the required test data themselves.

[0050] (4) This invention proposes a configuration process for swarm intelligence optimization algorithms, including population generation, population update, constraints and termination conditions. According to this process, researchers can configure personalized swarm intelligence optimization algorithms that meet the problem settings. Attached Figure Description

[0051] Figure 1 This is a flowchart of a benchmark testing method for optimizing collaborative manufacturing services on a network platform, according to an embodiment of the present invention.

[0052] Figure 2 This is a benchmark test system architecture diagram for network platform manufacturing service collaboration optimization according to an embodiment of the present invention.

[0053] Figure 3 The present invention relates to a flowchart for generating test data. Detailed Implementation

[0054] The present invention will now be described in further detail with reference to the accompanying drawings.

[0055] like Figure 1 As shown, the benchmark testing method for network platform manufacturing service collaboration optimization according to the present invention includes the following steps:

[0056] Step 1: Analyze the relationships between the elements of collaborative optimization of manufacturing services on the network platform, and construct a complex network model for collaborative optimization of manufacturing services on the network platform. The specific implementation method is as follows:

[0057] (1) For the manufacturing task network T Net = <T,sT,E T> where T represents a complex manufacturing task, sT represents a subtask, and E T This indicates the relationships between subtasks. The description of manufacturing task T includes the number of subtasks of task T. Mission arrival time Task deadline Task cost constraints Task quality constraints Task progress status Task completion status Task value and task completion assessment

[0058] The description of subtask sT includes the subtask type. Subtask deadline Subtask cost constraints Subtask quality constraints Subtask process constraints Subtask progress status Subtask completion status Subtask value Subtask completion assessment

[0059] Relationship between subtasks E T for in, Refers to the process relationship between subtasks. Subtask attribution;

[0060] (2) For manufacturing service network S Net = <S,E S >, where S represents manufacturing services, E S This indicates the relationships between services. The description of the manufacturing service S includes the type of service S. Service Costs Service quality Service efficiency Service capacity Service collaboration efficiency Service reliability Service failure status Service load status Service utility

[0061] Service Relationship E S for in, Refers to service similarity, Indicates the intensity of service collaboration. This refers to the service attribution relationship;

[0062] (3) The collaborative relationship between manufacturing services and tasks is represented as follows: Service S i Subtask sT was executed j Based on the collaborative relationship between manufacturing services and tasks Get the output of the service As shown in equation (1), the profit of the service As shown in equation (2), where, Service S i Execute subtasks sT j Cost:

[0063]

[0064]

[0065] Step 2: Construct a network platform manufacturing service collaboration optimization problem model. The specific implementation method is as follows:

[0066] (1) For the model configuration of the network platform manufacturing service collaboration optimization problem, determine the information scope, scheduling timing, scheduling scope and optimization objective of the collaboration optimization problem respectively;

[0067] First, the information scope of this collaborative optimization problem is defined, which includes complete information, partial information, and complete response. Complete information means that the entire set of tasks is known to the scheduling system from the beginning. Partial information means that the scheduling system can only process a portion of the tasks that can be foreseen. Complete response means that the scheduling system can only know the information of the task until it arrives.

[0068] Secondly, the scheduling timing for this collaborative optimization problem needs to be clarified. The scheduling timing includes periodic-driven scheduling and event-driven scheduling. The periodic-driven scheduling refers to scheduling once every fixed time period, while the event-driven scheduling refers to scheduling once whenever a task arrives.

[0069] Subsequently, the scheduling scope of the collaborative optimization problem is determined, which includes local scheduling and global scheduling. Local scheduling refers to selecting only a portion of tasks and services for scheduling in a single scheduling process, while global scheduling refers to considering all tasks and services in a single scheduling process and performing a global solution.

[0070] Finally, the scheduling system determines the scheduling objectives, which are divided into single-objective scheduling and multi-objective scheduling based on the number of objectives.

[0071] (2) Performance evaluation of network platform manufacturing service collaboration optimization refers to the evaluation of the overall system operation, which is divided into internal attributes and external attributes for separate evaluation;

[0072] First, the intrinsic attributes are evaluated, considering the service clustering coefficient of the manufacturing service network, the network throughput, and the network degree distribution; among these... Representative Service S i With S j similarity, Representative Service S i With S j complementarity, Representative Service S i With S j Based on the attribution relationship, the network degree distribution is divided into service similarity distribution. Service complementarity distribution and service attribution distribution The specific calculation methods are shown in equations (3)-(5):

[0073]

[0074]

[0075]

[0076] Secondly, external attributes are evaluated, taking into account system stability, load balancing, system robustness, and system reliability.

[0077] Step 3: Generate collaborative optimization test data for network platform manufacturing services, such as... Figure 3 As shown, the specific implementation method is as follows:

[0078] (1) Determine whether the collaborative optimization problem has real data and whether test data is needed. If test data is not needed, import the real data and proceed to the next test.

[0079] (2) Generate test data. First, describe the data related to the test problem. According to different types of test data, set the range and distribution of the corresponding test data to ensure that the test data conforms to the description of the problem during the generation process. Generate data for the service network and the task network respectively.

[0080] (3) For the service network, firstly, generate the service network scale, including the number of services and the number of enterprises; for the enterprise scale, if there is a demand, generate the scale of specific enterprises, including the number of enterprise services, the number of enterprise service clusters, and the capacity of enterprise service clusters, and check whether there is a conflict with the cluster distribution. If there is a conflict, repeat the above service network scale generation process until there is no conflict; then generate service clusters, including the number of service clusters and the service cluster capacity; for enterprise service attributes, if there is a demand, generate the service attributes of specific enterprises, including the number of enterprise service functions, enterprise service cost, enterprise service quality, enterprise service reliability, enterprise service efficiency, and enterprise service collaboration efficiency, and check whether they conflict with the service attribute distribution. If there is a conflict, repeat the above service attribute generation process until there is no conflict; then generate the remaining service attributes that meet the service requirements, including the number of service functions, service cost, service quality, service reliability, service efficiency, and service collaboration efficiency; finally, check whether there are any abnormal services and make modifications.

[0081] (4) For the task network, the first step is to determine whether there are requirements for enterprise task load. If there are requirements, generate enterprise task load, including enterprise task arrival, number of enterprise subtasks, and enterprise subtask size, and check whether they conflict with the task load distribution. If so, repeat the above task load generation process until there is no conflict. Then generate the remaining task load, including task arrival, number of subtasks, and subtask size. For enterprise task attributes, if there are requirements, generate task attributes for specific enterprises, including enterprise task chain length, enterprise task waiting tolerance, enterprise task delay tolerance, enterprise task quality tolerance, and enterprise task price tolerance, and check whether they conflict with the task attribute distribution. If there is a conflict, repeat the above task attribute generation process until there is no conflict. Then generate the remaining task attributes that meet the task requirements, including enterprise task chain length, task waiting tolerance, task delay tolerance, task quality tolerance, and task price tolerance. Finally, check whether there are any abnormal tasks and make modifications.

[0082] Step 4: Configure the network platform manufacturing service collaboration optimization test algorithm. The specific implementation method is as follows:

[0083] (1) Determine whether the collaborative optimization problem requires testing a specific external algorithm or configuring an existing algorithm. If no existing algorithm needs to be configured, import the specific external algorithm and proceed to the next test.

[0084] (2) For the configuration of the intelligent optimization algorithm, the first step is population generation; the population generation includes individual encoding to form a group encoding sequence; each individual in the group encoding sequence contains all the tasks that need to be scheduled, and the encoding sequence of each task contains the schemes of all subtasks, including the subtask start time and subtask scheduling service; then the size of the population is determined.

[0085] (3) The population update mechanism adopts swarm intelligence optimization algorithms, including bee colony algorithm, particle swarm algorithm, and evolutionary algorithm; users adjust the number of iterations and cross-evolution factors of the optimization algorithm to obtain personalized algorithms for testing;

[0086] (4) Set the constraints and termination conditions of the intelligent optimization algorithm. The user sets the range in which the intelligent optimization algorithm searches, the number of search iterations, and the final solution accuracy.

[0087] Step 5: Evaluate the results of the collaborative optimization test for the storage network platform manufacturing services. The specific implementation method is as follows:

[0088] Based on the manufacturing service collaboration optimization problem model to be tested, the test data, and the test algorithm, the collaboration optimization problem is solved. Users test its performance under different datasets and different algorithms according to different problem models, so as to make an intuitive comparison of the network platform manufacturing service collaboration optimization problem and solution under different backgrounds. After the solution is completed, the historical records are stored and retained, recording the most suitable test data and the optimal intelligent optimization algorithm configuration.

[0089] like Figure 2 As shown, a benchmark testing system for network platform manufacturing service collaboration optimization according to the present invention includes: a manufacturing service collaboration benchmark testing system model library, a database, and an algorithm library. The system configures a model based on the manufacturing service collaboration optimization problem to be tested, stores the model in the model library, then retrieves test data adapted to the model from the database, and then retrieves test algorithms adapted to the model from the algorithm library to solve the collaboration optimization problem. The results are then compared and displayed on the benchmark testing system. Simultaneously, the algorithm library stores the test comparison results in the database, and the database provides feedback and corrections to the model based on the quality of the test results.

[0090] In summary, this invention discloses a benchmark testing method and system for network platform manufacturing service collaboration optimization. This invention effectively addresses the lack of a unified descriptive model, test data generation, and algorithm configuration method in the field of network platform manufacturing service collaboration optimization, enabling horizontal comparative evaluation across related research areas. The method includes five steps: analyzing the relationships between elements in network platform manufacturing service collaboration optimization; constructing a complex network model for network platform manufacturing service collaboration optimization; configuring the network platform manufacturing service collaboration optimization problem model as needed; importing or generating test data as needed; importing or configuring algorithms for solving the problem as needed; solving the test problem multiple times; and recording and storing the test results. Therefore, this invention enables unified modeling and problem description for network platform manufacturing service collaboration optimization, generates test data according to a standardized process, and configures algorithms, effectively allowing for unified comparison of various current studies on network platform manufacturing service collaboration optimization. This is of great significance for research management capabilities in this field and for efficient and stable production in the manufacturing sector.

[0091] The contents not described in detail in this specification are existing technologies known to those skilled in the art.

[0092] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A benchmarking method for network platform manufacturing service collaboration optimization, characterized in that, The method comprises the following steps: Step 1: analyzing the relationship between the network platform manufacturing service cooperation optimization elements, and constructing a network platform manufacturing service cooperation optimization complex network model; in the process of network platform manufacturing service cooperation optimization, including providers, demanders and managers, from the perspective of complex network model construction, the model elements are divided into manufacturing service resources, manufacturing service demands and the association relationship between the supply and demand, and the network platform manufacturing service cooperation optimization complex network model is constructed according to the model elements; Step 2: configuring the network platform manufacturing service cooperation optimization problem model according to the actual problem; the problem configuration and performance evaluation analysis are carried out, wherein the problem configuration refers to setting various assumptions and constraint conditions of the optimization problem, and the performance evaluation refers to evaluating the overall cooperation process on the basis of cooperation optimization; Step 3: considering the demand of the test problem, importing or generating test data as needed; first, it is judged whether the demand of the test problem has real data, if it has real data, the real data is directly imported for testing; If there is no real data, the test data is simulated and generated; in the case that various data do not conflict with each other, the test data meeting the demand of the test problem is generated; Step 4: considering the test algorithm, importing or configuring the test algorithm for solving the test problem as needed; first, it is judged whether the test problem has the demand of the test specific algorithm, if it has the demand, the specific algorithm corresponding to the demand is imported; if there is no demand of the test specific algorithm, the algorithm is configured as needed according to the model of the test problem; the algorithm configured as needed is an intelligent optimization algorithm; Step 5: solving the test problem multiple times, and recording and storing the test results of the test problem; considering the random uncertainty of the test problem, if the test object is a problem model, different test data and solving algorithms are used for solving to obtain an approximate optimal solution; if the test object is an algorithm, different test data and other algorithms are used for comparison to obtain the best comparison effect; the other algorithms include multi-swarm intelligent algorithms and improved evolutionary algorithms.

2. The benchmarking method for collaboration optimization of network platform manufacturing services as claimed in claim 1 wherein: The step 1 specifically comprises: (1) For manufacturing task network wherein, represents a complex manufacturing task, represents a subtask, represents an association relationship between subtasks, for manufacturing task description includes task subtask quantity , task arrival time , task deadline , task cost constraint , task quality constraint , task progress status , task completion status , task value , and task completion evaluation ; Description of subtasks includes subtask type , subtask deadline , subtask cost constraint , subtask quality constraint , subtask procedure constraint , subtask progress status , subtask completion status , subtask value , subtask completion evaluation ; Correlation between subtasks For Wherein, Indicating subtask procedure relationship, Indicating subtask attribution relationship; (2) For manufacturing service networks ,in, Indicates manufacturing services. This indicates the relationships between services, for example, in manufacturing services. The description includes services Type Service costs Service quality Service efficiency Service capacity Service collaboration efficiency Service reliability Service unavailable Service load status ; Inter-service association relationship For Wherein, Indicating service similarity, Indicating service cooperation strength, Indicating service belonging relationship; (3) The cooperation relationship between manufacturing services and tasks is represented as , representing the service executes the subtask ; the output of the service is obtained according to the cooperation relationship between manufacturing services and tasks The profit of the service is shown as formula (1) The profit of the service is shown as formula (2), wherein is the cost of the service executing the subtask :​ （1） （2）。 3. The benchmarking method of collaborative optimization of network platform manufacturing services as claimed in claim 2, wherein: The step 2 specifically comprises: (1) for the network platform manufacturing service cooperation optimization problem model configuration, respectively determine the information range, scheduling time, scheduling range and optimization target of the cooperation optimization problem; First, the information range of the cooperation optimization problem is defined, and the information range includes complete information, partial information and complete reaction; the complete information refers to that all task sets are known by the scheduling system at the beginning, the partial information refers to that the scheduling system can only handle a predictable part of the task, and the complete reaction refers to that the scheduling system can only know the information of the task when the task arrives; Secondly, the scheduling time of the cooperation optimization problem is determined, and the scheduling time includes periodic driving scheduling and event driving scheduling; wherein, the periodic driving scheduling refers to scheduling once every fixed time period, and the event driving scheduling refers to scheduling once when a task arrives; Subsequently, the scheduling range of the collaborative optimization problem is determined, and the scheduling range includes local scheduling and global scheduling; wherein the local scheduling refers to selecting a part of tasks and services for scheduling in one scheduling process, and the global scheduling refers to considering all tasks and services in one scheduling process and performing global solving; Finally, the scheduling system determines the scheduling target, and according to the number of the scheduling target, the scheduling is divided into single-target scheduling and multi-target scheduling; (2) The performance evaluation of the network platform manufacturing service collaborative optimization refers to the evaluation of the overall operation of the system, which is divided into internal attributes and external attributes for separate evaluation; Firstly, the intrinsic attributes are evaluated, including the clustering coefficient of service community of manufacturing service network, the network flow of manufacturing service network and the network degree distribution; wherein, representative service with the similarity, representative service with the complementary degree, representative service with the attribution, thus, the network degree distribution is divided into service similarity distribution , service complementary degree distribution and service attribution distribution ; the specific calculation method is shown in equations (3)-(5): （3） （4） （5） Secondly, the external attributes are evaluated, and the system stability, load balancing, system robustness and system reliability are considered.

4. The benchmarking method of collaborative optimization of network platform manufacturing services as claimed in claim 3, wherein: The step 3 specifically includes: (1) It is determined whether the collaborative optimization problem has real data and whether test data is needed. If the test data is not needed, the real data is imported, and the next step of testing is performed; (2) The test data is generated. First, the test problem related data is described, and according to different types of test data, the range and distribution of the corresponding test data are set to ensure that the test data generated conforms to the description of the problem, and the data of the service network and the task network are generated respectively; (3) For the service network, first generate the service network scale, including the number of services and the number of enterprises; for the enterprise scale, if needed, generate the scale of specific enterprises respectively, including the number of enterprise services, the number of enterprise service groups, the capacity of enterprise service groups, and check whether there is a conflict with the group distribution. If there is a conflict, repeat the above service network scale generation process until there is no conflict; then generate the service group, including the number of service groups and the capacity of service groups; for the service attribute of the enterprise, if needed, generate the service attribute of the specific enterprise respectively, including the number of enterprise service functions, the cost of enterprise services, the quality of enterprise services, the reliability of enterprise services, the efficiency of enterprise services, the collaboration efficiency of enterprise services, and check whether they conflict with the service attribute distribution. If there is a conflict, repeat the above service attribute generation process until there is no conflict; then generate the remaining service attributes that meet the service demand, including the number of service functions, the cost of services, the quality of services, the reliability of services, the efficiency of services and the collaboration efficiency of services; finally, check whether there is an abnormal service and modify it. (4) For the task network, the first step is to determine whether the enterprise task load has requirements, and if there are requirements, to generate the enterprise task load, including enterprise task arrival, enterprise sub-task quantity, enterprise sub-task size, and check whether they conflict with the task load distribution, if there is, repeat the above task load generation process until there is no conflict; Then generate the remaining task load, including task arrival, sub-task quantity and sub-task size; For enterprise task attributes, if there are requirements, generate specific enterprise task attributes respectively, including enterprise task chain length, enterprise task waiting tolerance, enterprise task delay tolerance, enterprise task quality tolerance, enterprise task price tolerance, and check whether they conflict with the task attribute distribution, if there is, repeat the above task attribute generation process until there is no conflict; Then generate the remaining task attributes that meet the task requirements, including enterprise task chain length, task waiting tolerance, task delay tolerance, task quality tolerance, and task price tolerance; Finally, check whether there are abnormal tasks and make modifications.

5. The benchmarking method of collaborative optimization of network platform manufacturing services as claimed in claim 4, wherein: The step 4 comprises: (1) Determine whether the specific external algorithm needs to be tested for the collaboration optimization problem, whether the existing algorithm needs to be configured, and if not, import the specific external algorithm and proceed to the next step of testing; (2) For the configuration of intelligent optimization algorithm, the first step is population generation; The population generation includes individual coding, forming a group coding sequence; Each individual in the group coding sequence contains all the tasks that need to be scheduled, and each task coding sequence contains the scheme of all sub-tasks, including sub-task start time and sub-task scheduling service; Then determine the size of the population; (3) The population update mechanism adopts swarm intelligence optimization algorithm, including bee colony algorithm, particle swarm algorithm, and evolution algorithm; The user adjusts the number of iterations of the optimization algorithm and the cross-evolution factor to obtain a personalized algorithm for testing; (4) Set the constraint conditions and termination conditions of the intelligent optimization algorithm, and the user sets the search interval, search iteration number, and final solution accuracy of the intelligent optimization algorithm.

6. The benchmarking method of collaborative optimization of network platform manufacturing services as claimed in claim 5, wherein: The step 5 comprises: According to the manufacturing service collaboration optimization problem model to be tested, the test data and the test algorithm, the collaboration optimization problem is solved; The user tests the performance of different problem models under different data sets and different algorithms to intuitively compare the network platform manufacturing service collaboration optimization problem and the solution method under different backgrounds; When the solution is completed, the historical records are stored and retained, and the most suitable test data and the optimal intelligent optimization algorithm configuration are recorded.

7. The system for benchmarking the collaboration optimization of network platform manufacturing services according to one of claims 1 to 6, characterized in that Comprise: The model library, the database and the algorithm library of the manufacturing service cooperation benchmark test system are manufactured, the model is configured according to the manufacturing service cooperation optimization problem to be tested, the model is stored in the model library, then, the test data suitable for the model is called from the database, the test algorithm suitable for the model is called from the algorithm library, the cooperation optimization problem is solved, and the result comparison and display are carried out on the benchmark test system; meanwhile, the test comparison result is stored in the database by the algorithm library, and the model is feedback corrected according to the advantages and disadvantages of the test result by the database.

Images