Dynamic Self-adaptation Approach Of Cloud Manufacturing Service Composition Based On Complex Network

Cloud manufacturing,as a new practical paradigm that enables the servitization and intellectualization of manufacturing industry,has received extensive attentions from academia and industry.In cloud manufacturing,manufacturing service composition is an important means to achieve synergy,integration,sharing and value-adding of cross-domain and heterogeneous manufacturing resources.By composing multiple manufacturing services with simple function on demand,composite service can be constructed flexibly to fulfill the complex and varied manufacturing requirements of users.However,the complexity and dynamic of manufacturing requirements and manufacturing resources make the availability,reliability,and credibility of manufacturing service composition difficult to guarantee.How to improve the self-adaptation ability of manufacturing service composition in the dynamic and complex cloud manufacturing environment has become the key issue to be solved urgently.Based on the complex network theory,taking enterprises in industry alliance as the application object,this dissertation conducts in-depth research on the cloud manufacturing service composition dynamic self-adaptation approach.The main research work includes:(1)Cloud manufacturing service complex networks modeling for industry allianceAiming at the complexity and dynamic of manufacturing tasks and manufacturing services,the cloud manufacturing service complex networks modeling approach for industry alliance is proposed.The connotation and characteristics of industry alliance are analyzed.Then the cloud manufacturing for industry alliance(CMfgIA)which is a new application mode for cloud manufacturing is proposed,and its conceptual model and operation framework are designed.Based on the domain engineering and multi-granularity theory,the multi-granularity manufacturing requirement feature models and manufacturing service mapping mechanism are constructed.It realizes the granularized modeling,hierarchical management and diversified reuse of manufacturing requirements and manufacturing services,improves the domain completeness of manufacturing requirements and manufacturing services,and thus ensures the controllability of dynamic changes as well as the availability of manufacturing service composition.Finally,the cloud manufacturing service complex networks modeling approach is expounded,and the manufacturing tasks complex network and manufacturing services complex network are constructed,which provides theoretical support for the self-adaptation of cloud manufacturing service composition in the complex and dynamic environment.(2)Cloud manufacturing business process decomposition approach based on complex networkAiming at the problem that the manufacturing requirement feature model is difficult to construct because of the high complexity of cloud manufacturing business process,a cloud manufacturing business process decomposition approach based on complex network is put forward.The complexity of cloud manufacturing business process requirements in industry alliance is analyzed.Based on the logical dependency relationships of manufacturing business activities in the domain,the business process complex network is constructed.Then the business process complex network decomposition(BPCND)approach is presented.Its decomposition mechanism is elaborated and the corresponding decomposition algorithm is designed.In the algorithm,the weight of structural similarity and domain similarity can be dynamically adjusted during the decomposition process through an adaptive optimization strategy.And the optimal decomposition results can be obtained without the pre-given optimal number of sub-processes and weight information.Through the experiment of elevator design business process decomposition,the validity and superiority of the decomposition approach are verified from four indicators which are Modularity,Density,Entropy and TDiff.Therefore,it can provide approach support for improving the modeling efficiency and scientificity of the complex manufacturing requirement model.(3)Manufacturing service composition self-adaptation approach based on dynamic matching networksAiming at the problem that the reliability and credibility of manufacturing service composition in dynamic environment is difficult to guarantee,a manufacturing service composition self-adaptation approach based on dynamic matching networks is proposed.The manufacturing task-manufacturing service dynamic matching networks(DMN)is established.Based on DMN,the self-adaptation mechanism of manufacturing service composition is designed,and the three-phase self-adaptation process,namely,dynamic perception-dynamic evaluation-dynamic reconstruction,is expounded.Then the self-adaptation model of manufacturing service composition is presented.Driven by the perception of the dynamic changes of manufacturing service composition,the dynamic update of manufacturing tasks complex network and manufacturing services complex network is realized.Meanwhile,the self-adaptation algorithm for manufacturing service composition is designed.In the algorithm,the manufacturing service load and dynamic QoS which is evaluated by a load queue model are applied as the optimization objects,and the optimal manufacturing service composition problem is transformed into the performance optimization of the manufacturing services complex network.And thus the dynamic self-adaptation of manufacturing service composition is realized.Through the simulation experiment of elevator design service composition,the algorithm is proved to have good adaptability to dynamic environments from four indicators which are execution success rate of manufacturing task,load balancing of manufacturing service cluster,global optimization capability for load queue,average QoS utility of manufacturing service composition.(4)Application of cloud manufacturing service composition self-adaptation in the elevator industry allianceThe application of cloud manufacturing service composition self-adaptation in the elevator industry alliance is studied.The characteristics of the elevator industry alliance are analyzed.Then a cloud manufacturing service platform for the elevator industry alliance is designed and constructed.Taking the elevator design and calculation as an example,the elevator manufacturing service modeling and management,as well as the elevator manufacturing service composition selfadaptation are researched.The customization and self-adaptation of the elevator design and calculation cloud service system is realized.It can meet the dynamic and complex design business requirements of the enterprises in the elevator industry alliance.Therefore,the feasibility and effectiveness of the proposed theory and approaches are verified.