Application of an improved Hungarian algorithm in semantic WEB service discovery

The important step towards Web service implement is the discovery of required services. Traditional Web service uses UDDI, WSDL standards etc. Keywords based searching is used in the traditional service discovery. For needing manual intervention and lack of semantic description, the methods before have low precision rate. Semantic Web service problem can be reduced to an assignment problem. A mathematical model is achieved to account for the assignment problem. Similarity matrix is formed based on service request interface parameters and service advertisement interface parameters. The improved Hungarian algorithm is described to find the answer of mathematical model and discover the optimum service by searching for feasible matching plan. Meanwhile an algorithmic instance is given to illustrate the improved Hungarian algorithm could help requester find out the right request service. The test indicates that this algorithm increases the precision ratio and the recall ratio of Web service.