Based On The Agent Of Semantic Web Services Middleware Design

With the development of the Internet, various new applications are emerging. Avariety of sensors and mobile devices are integrated into the Internet, which is theso-called Internet of Things. The primary problem is to establish unified accessinterfaces through which users can access and manage these devices using technologiesof the Internet. Secondly, the problem is how to build application relations between thehuman and things, in other words, how to discover and integrate service information tosatisfy task requirements of human beings. For these issues, this paper employssemantic technology to build service and ontology data model for sensors and theirowners. We also designs a multi-agent system based on these semantic service.This paper first studies the framework and defines two agent structures: the middleagent and management agent. The management agent is an abstraction modeling humanactivity and the middle-agent is an abstraction modeling sensor devices in the Internetof Things. They form a hybrid multi-agent system which implements the interactionsbetween human beings and things.Secondly, we design the structure of the management agent based on work-flowtask model and propose a heuristic service discovery algorithm based on multi-agentcombined information evaluation. The algorithm employ combined table to recordservice combined information and neighbors could exchange combination to getinformation of each other. Besides, the algorithm evaluate service request relying oninformation about agents’ neighbors and select the most relevant neighbors to forwardthe request information.On the other hand, we design middle agent structure based on agent-basedsemantic services matching algorithm. This algorithm determines the rules anddependencies of executing semantic service by matching and reasoning based on middleagents’ local knowledge model. Finally, we summarize this work by analyzing itsachievements and shortcomings and describe the future work.