The Web Service Matching Algorithm Based On Semantic And Qos Research

With the rapid development of Internet, web services have been widely used. It hasevolved from simple web pages into a wide variety of network resources, and the amoutof web services available on the Intemet is becoming larger and larger, and more andmore complex. In addition, the demands of service requesters on web services are getetting higher and higher. They hope that the selected services not only meet theirnecessary functional requirements, but also are with fast response and strongavailability. They start to pay attention on non-functional attributes of services, mainlyon the quality of service. Currently, most web services matching are based on keyword,and cannot support semantic matching. These will lead to return results with low recalland precision. Moreover, they rarely take the requirements of quality into account. Fewhave done, but only calculate the values of quality parameters simply, and donn’tcombines them with the service requesters’ actual requirements. So the selected webservices are often not the closest to the needs of the service requester. In-depth study ofweb service and its matching method, we take semantic and QoS (Quality of Service)technology into the matching of web services, and discuss the QoS ontology model,web services matching process based on semantic and QoS, and matching algorithms indetail. The main work is as follows:Firstly, semantic description of web services and QoS models are researched. Weimprove and perfect the hierarchical structure and metrics of current QoS ontologymodel, so that it can support the exact metric, interval metric and linguistic metric andnumeric value, Boolean value and linguistic value, so as to provide the basis for thefollowing semantic matching and QoS parameters matching.Secondly,a multi-level and fine-grained QoS-based semantic web service matchingprocess is described in detail. And then focus on the existing semantic matchingalgorithm and its shortcomings. Based on it, we propose an improved similarityalgorithm based on semantic distance, and the weight distribution method based on thedepth, density and relation between concepts, improve the GCSM distance formula andmodify the traditional similarity function, so that it can be more accurate and flexible to calculate ontology similarity, and improve the recall rate and precision rate.Thirdly，due to it’s difficult to better meet the QoS requirements of the servicerequestors, we improve and perfect the existing QoS matching process and algorithm tomake it meet the functional requirements and support select more appropriate servicewhich is more closer to the QoS requirements of requesters.At last，we simulate the similarity algorithm based on matching process andalgorithms to verify the correctness and feasibility. And in order to reflect thesuperiority of the proposed algorithm, we compare it with the traditional algorithm.