Research On Resource Service Search In Cloud Manufacturing Base On Swarm Intelligence

At present, China’s manufacturing industry experiences a transition from theproduction model to the service model, which is becoming a new economic growthpoint. However, realizing this transition need related technologies, platforms and anew model. With the development of cloud computing, networking and othernext-generation IT technologies, cloud manufacturing model emerges with mixing anumber of advanced manufacturing technologies and modes. Cloud Manufacturingaims to gather global resources to achieve the resources optimal allocation andprovides a core power for speeding up transformation and upgrading of traditionalmanufacturing economy.As one of the core technologies of cloud manufacturing, services searching is thekey technology to achieve resources optimal allocation and on-demand use. Servicessearching and discovery provides data for the realization of the services compositionand optimization, which performance impacts the optimal allocation of resourceservices directly. Whereas, most of the available services searching researches focuson the computing resources and web services, few studies on services searching in themanufacturing field. Due to the heterogeneity and complexity of manufacturingresources, the existing services searching methods can not work well in themanufacturing field. Consequently, this paper presents a system framework ofmanufacturing resources searching through combing the methods of web servicesdiscovery and the characteristics of manufacturing resources. Then, we present aQoS-based services searching model based on the research of the virtualization andservitization of manufacturing resources. Finally, we have done some research on theservice searching base on Bee algorithm.In the first part, we propose a QoS-based service searching system architecture,consisting of three links for realizing services searching, service extraction, ssessmentand comparison, based on the existing service composition and optimization models,key technologies and services processes. This system gives the technical accurate routes and workflows to achieve resources discovery and provides basic informationfor matching of cloud manufacturing services.Secondly, with the deficiency of the way for services description, such asinsufficient service description information and unclear classification, the paperanalyzes the features of cloud manufacturing resources and uses OWL to describeresource basic information, manufacturing function information and QoS information.According to the users’ requirement of service cost-performance, we purpose the QoSevaluation indexes and calculation formulas and establish a quaternary servicediscovery model with obligatory targets.As the bees algorithm on convergence and stability are superior to the traditionalswarm intelligence optimization algorithm, and can find the optimal solution fastercomplex optimization problems, and is robust, easy to combine with other methods,excellent advantages of distributed calculation mechanism. In this paper, an enhancedalgorithm as a cloud manufacturing resource service search algorithm is used.Finally, based on these studies, we build a QoS-based cloud service discoverysimulation testing environment. The experiment using improved Bees algorithmshows that the convergence time and convergence precision of search algorithmimproved a lot, compared with the original Bees algorithm.