Research On Cloud Manufacturing Resource Allocation Method Based On Complex Network

With the advent of a new round of information revolution represented by intelligence,the traditional manufacturing is in urgent need of transformation and upgrading to adapt to the changing market environment.It has become a new trend for manufacturing industry to share various manufacturing service resources to meet the customized needs of customers with the support of network,in which the efficient and rational allocation of cloud manufacturing resources is its core issue.Focusing on the problems of limited accuracy and less collective profit in the process of configuring cloud service resources,on the basis of building the resources matching framework for cloud tasks,this thesis studied the decomposition and aggregation of complex cloud tasks,the matching between tasks and resources and optimization of service resources based on collective interests through complex networks and evolutionary game theory.The main research contents are as follows:The cloud task decomposition and aggregation method based on complex network was proposed.The basic unit of cloud task decomposition was determined,and the specific content and steps of complex task decomposition were presented.For the complex cloud task,the measurement index of task granularity was proposed,and the basic principles and related constraints of task decomposition were given.According to the task relevance between the meta-tasks,the association undirected graph was established and segmented by using graph group detection algorithm,which achieved the purpose of complex task decomposition and aggregation.The service resource screening and matching method based on complex network was proposed.According to the task constraint structure,a comprehensive semantic manner was taken to describe the sub-tasks obtained from the complex cloud manufacturing task decomposition and aggregation.Based on the two-mode network,the task-resource network model was constructed.Based on the fuzzy comprehensive evaluation of the service resources,the task requirements were taken as the reference sequence and the matching degree between the candidate service resources and the tasks was gained by the grey correlation analysis method.The two-mode network community discovery algorithm was used to obtain the resource candidate pool of each sub-task,and the quality of the community division was evaluated by the modular function,which ensured the rapid and reasonable screening and matching of resourcesThe service resource optimization model based on the collective interests of cloud manufacturing alliances was constructed.Based on the sub-task candidate service resource set,the evolutionary game network model of task issuer side and service resource side was established.Through the foreground theory,an incentive method considering the satisfaction of the estimated task completion was proposed,which expands the game strategy space and improves the task completion efficiency of the alliance.The Fermi update rule was introduced to bring randomness in the strategy learning process of each participating entity,which evolved the whole process of the game network reaching equilibrium state.With the goal of the largest collective profit of the cloud manufacturing alliance,the optimal service resource combination plan was chosen.The cloud manufacturing resource configuration prototype system was developed.Based on the analysis of the prototype system function modules and business processes,the IntelliJIDEA was used as the development platform,the front-end interface was designed by Vue framework,the related algorithms were finished by MATLAB,data storage was done by MySQL database,and the Java programming language was used to integrate the front interface and the server of back end.Finally,a resource allocation prototype system for complex cloud manufacturing tasks was completed,which verified the feasibility and effectiveness of the proposed model and method.