A Hybrid Matchmaking Algorithm for Semantic Web Service

An important component of the Web service is the matchmaking between service requests and service advertisements. This paper concentrates on the development of a hybrid Web service matchmaking approach that aims to combine the merits of fuzzy rough set theory and of bipartite graph matching. First, function elements, such as input properties and output properties are extracted from Services Profile of service advertisements and service requests which are expressed in terms of OWL-S semantic description, respectively. Then, an incomplete information system is described and converted into a simple system. In order to address uncertainty in service properties, the fuzzy rough set theory is adopted to eliminate redundant function properties. In addition, according to the remaining properties, a bipartite graph is constructed. Hence the calculation of similarity degree will be switched to the problem of extended optimal matching for bipartite graph. Finally, we present some matchmaking results of the bipartite graph matching algorithm and of the proposed hybrid algorithm and analyze the rationality of the proposed hybrid algorithm.