Agent-based Dynamic Service Composition In Cloud Computing

With the increase of number and kinds of web service,servicecomposition has become one of the key problems for service.How to selectservice and How to optimize the service composition must be taken intoconsideration when we start service composition.Traditional web servicetechnology is just a response without any statements.It has limitedfunctions,it can not response events initiative and the cooperation betweenservices can not be started initiative.In order to endow the web services withability of initiative a common choice is combine web service technology withagent technology for the features of agent,such as automation, sociality,reflection, initiative.As the composition of web services, agents can completeservice composition with high quality through communication with eachother and learn from each other.As the development of cloud computing, webservice change into cloud service step-by-step.Traditional method which beused in service composition cannot fulfill the requirement of cloud service.And other problems such as uniform modeling,mass cloud servicemust be taken into consideration.In order to solve the problems above, an agent model is proposed whichsuit the cloud computing environment. Besides, we also propose aoptimization method for the features of modeling in cloud computing.Through analysis the history data of service composition, reasoning thepreference of Customers and build decision tree to choose service,we cancomplete the service composition with high quality.