Research On Complex Task Decomposition And Resource Allocation In Cloud Manufacturing Environment

In recent years,the development of information technologies such as the Internet of Things,big data and cloud computing has had a profound impact on the transformation and upgrading of the manufacturing industry.In response to this change,China has put forward the "Made in China 2025" program,which focuses on high-tech complex manufacturing and promotes the "deep integration of information technology industry and manufacturing industry".As a new network manufacturing mode combining information technology and manufacturing,cloud manufacturing is essentially idle manufacturing resources to provide users with manufacturing services that meet their needs.Therefore,the realization of efficient and reasonable resource allocation is the core and key technology of cloud manufacturing platform.Cloud manufacturing platform provides services for complex manufacturing tasks,which needs to go through three stages: task decomposition,resource and task matching and resource allocation optimization.In accordance with this process,this paper studies the resource allocation of cloud manufacturing platform.In the stage of complex manufacturing task decomposition,a meta-task recombination optimization method based on ant colony clustering algorithm was proposed.This paper describes the correlation of meta-tasks from the perspectives of structure,function,information and logistics,and analyzes the calculation methods of meta-task correlation under the four structures of serial,parallel,cyclic and coupling.Ahp is used to assign weight to the importance of each index of meta-task correlation,and a matrix reflecting the comprehensive correlation of meta-tasks is obtained.On this basis,a meta-task recombination optimization model was established to maximize the cohesion,coupling degree and equilibrium degree of subtasks,and the decomposition problem of complex manufacturing tasks was transformed into a meta-task clustering recombination problem.Considering the inapplicability of conventional clustering algorithm,a decentralized ant colony clustering algorithm was designed to solve the model based on the basic principles of ant colony algorithm.The superiority of the proposed algorithm is verified by the analysis and comparison of examples.In the stage of resource and task matching,the knowledge graph technology is introduced and the construction method of resource and task matching knowledge graph is proposed.The quad structure of knowledge graph is analyzed,and a top-down knowledge graph construction method based on ontology model is proposed for resource and task information characteristics in cloud manufacturing environment.The ontology model of resources and tasks is constructed.By extracting the functional information of the ontology model,the matching information model of resources and tasks is further constructed.The matching rules of resources and tasks are designed and described based on SWRL.Taking manufacturing task as an example,Neo4 j software was used to construct the knowledge map of resource and task matching and display it visually.In the optimization stage of resource allocation,a multi-objective optimization model was constructed considering the interests of three parties involved in cloud manufacturing and an improved NSGA-Ⅲ algorithm was proposed.In order to guarantee the interests of all participants in cloud manufacturing,indexes reflecting the interests of all parties are selected and a multi-objective optimization model of maximizing the interests is established.In view of the disadvantage that the performance of NSGA-Ⅲ algorithm decreases with the increase of the target dimension,an adaptive evolutionary mechanism of multi-target population is proposed,which is combined with the reverse learning mechanism to improve the performance of NSGA-Ⅲ algorithm.The advantages of the improved NSGA-Ⅲalgorithm in convergence,diversity and robustness are verified by the test function.The results show that the fitness of the solution set of the improved NSGA-Ⅲ algorithm is better than that of the NSGA-Ⅲ algorithm in each index direction,and the candidate scheme of resource matching can be given which is more in line with the interests of all parties.By studying the complex task decomposition,resource and task matching and resource allocation optimization in the cloud system environment,the theoretical basis of the core key technologies of the cloud manufacturing platform is enriched,which is conducive to the realization of efficient and reasonable resource allocation.