A hybrid approach to semantic web services matchmaking

摘要

Deploying the semantics embedded in web services is a mandatory step in the automation of discovery, invocation and composition activities. The semantic annotation is the "add-on" to cope with the actual interoperability limitations and to assure a valid support to the interpretation of services capabilities. Nevertheless many issues have to be reached to support semantics in the web services and to guarantee accurate functionality descriptions. Early efforts address automatic matchmaking tasks, in order to find eligible advertised services which appropriately meet the consumer's demand. In the most of approaches, this activity is often entrusted to software agents, able to drive reasoning/planning activities, to discover the required service which can be single or composed of more atomic services. This paper presents a hybrid framework which achieves a fuzzy matchmaking of semantic web services. Central role is entrusted to task-oriented agents that, given a service request, interact to discover approximate reply, when no exact match occurs among the available web services. The matchmaking activity exploits a mathematical model, the fuzzy multiset to suitably represent the multi-granular information, enclosed into an OWLS-based description of a semantic web service.