Research On Dynamic Importance Evaluation And Adaptive Monitoring Method For Cloud Manufacturing Services

In cloud manufacturing mode,all kinds of manufacturing resources and capabilities are encapsulated into virtual manufacturing service,which can be fully shared and used on demand through the cloud platform.Enterprises can customize cloud manufacturing service system(CMSS)online through service composition to complete various manufacturing business activities.Due to the dynamic and unpredictability of the cloud environment,manufacturing services are uncertain,service failures and performance degradation often occur.As a result,it is difficult to meet users' customization and use requirements.Therefore,online monitoring of manufacturing services is usually required to provide a basis for ensuring the normal operation of CMSS.However,due to the large number of manufacturing services in cloud platform,continuous monitoring of all manufacturing services is expensive,poor operability and easy to cause resource waste.One possible solution is to prioritize the allocation of monitoring resources to the important manufacturing services.Therefore,how to accurately find a part of manufacturing service with high importance and then develop appropriate monitoring strategies is of great significance to rationally allocate resources and reduce maintenance costs on the premise of effective monitoring.To solve the above problems,based on complex network theory and evidence reasoning(ER)rule method,the research path of "modeling ?evaluation ? monitoring" is given,then a dynamic importance evaluation and adaptive monitoring approach for cloud manufacturing services importance is proposed.The specific research contents are as follows:(1)Domain-oriented cloud manufacturing services network modeling and evolution.The principle of a domain-oriented CMSS customization method is presented.Then the cloud manufacturing services complex(CMSCN)network modeling method based on collaborative interaction relationship of services in the domain is proposed.In the CMSCN,the function-oriented business process network and non-function-oriented service instance network are constructed respectively.After that,the BBV(Barrat-Barthelemy-Vespignani)model is apply to analyze the evolution process of the constructed CMSCN model,and the dynamic description of manufacturing service status and relationship is realized.(2)Cloud manufacturing service importance evaluation based on network features and evidential reasoning rules.Based on the contructed CMSCN,the dynamic importance evaluation mechanism for manufacturing services is described.Based the idea of on multi-index evaluation,an importance evaluation method for cloud manufacturing services based on centricity evidence is put forward.In the method,different centricity indexes are transformed into evaluation evidence and the calculation method of reliability and weight of the evidence is introduced.Then the ER rule is used to fuse the evaluation evidence and the belief distribution of importance of manufacturing services.Based on this,the final importance ranking results can be obtained.Finally,the effectiveness and superiority of the proposed method are verified by the importance evaluation experiments of vertical elevator design cloud services.(3)Adaptive monitoring method for cloud manufacturing services based on dynamic importance.According to the evaluation results and variation characteristics of functional importance and non-functional importance of manufacturing service,an adaptive monitoring strategy for cloud manufacturing service based on dynamic importance is proposed.Then the prototype platform for adaptive monitoring of vertical elevator design cloud service is built,and the corresponding monitoring function module is designed and developed.Finally,monitoring experiments are conducted based on the platform,and the feasibility and effectiveness of the proposed monitoring strategy are verified.