The Research And Realization Of Trust Degree Model Of Semantic Web Service Composition Based On Workflow Model

In recent years, people requirement for web services become more and more complex, so web service composition is welcomed by users. It can not only improve the reuse rate of service components, reduce the cost of development, can also compose flexibly according to user's demand. Service composition fast and accurately has become a huge challenge. The function and performance of component services is an important reference for component service discovery and service composition optimization. The semantic web technology unify the web service function description, it can help machines to make clear the meaning of service function, but the effectiveness of the component services function description cannot be guaranteed. Quality of service attribute values are the service performance evaluation standard, but the authenticity of the value of the Qo S attribute can not be confirmed. In the actual environment, because of the pressure of competition, service developers may exaggerate Qo S attribute values or provide false information to improve the probability function description. Therefore, the false information directly affects the choice of component services, so that users make a wrong judgment.In order to solve this problem, a trust model of semantic web service composition is established, which quantitative the users' trust of semantic web service composition separately from two aspects of the performance and function, and it provides a good method for the component service discovery and the service composition optimization. Semantic web service composition trust degree mainly includes two types of non-functional attribute trust degree and functional attribute trust degree. In the process of computing the non-functional attributes trust degree, improved the previous single objective reference calculation method which refers records or subjective user evaluation, combines the advantages of them. Use subjective user evaluation to calculate the trust degree of unrecorded type non-functional attributes, and analysis objective running record for calculating the trust degree of recorded type non-functional attributes, it can improve the calculation accuracy of trust degree of non-functional attributes. In the process of computing the functional attributes trust degree, considered the similarity of semantic functional description between the candidate components and the abstract nodes in the web service, and also take into account the matching degree of functional interface between the predecessor component and the following component. Horizontal and vertical comparison can ensure that functional attribute trust degree of component web service can comprehensively reflect the reliability of the component web service.A semantic Petri net model based on trust degree is proposed, which provides a formal modeling tool for semantic web service composition process, and implements the semantic web component service discovery mechanism based on trust model. Compared to the traditional Petri net model, the semantic Petri net model based on trust degree adds the semantic description information of transition sets for calculating semantic functional similarity degree between candidate components service and abstracted nodes, while adding related information matrix of the interface library collection for calculating the matching degree of functional interfaces of the candidate components. Through functional semantic description similar degree and functional interface attribute matching degree, we can obtain the trust degree of candidate component services, component service research device refers semantic web services or semantic web service composition function attribute of trust degree to found semantic web services or semantic web service composition process fragment which meets user functional requirements for generating candidate components service set of abstract node of the service composition. According to the component services discovery method based on function attribute trust degree of semantic web service composition, can get candidate component service collections for each abstract service nodes which include similar function. So the process needed to optimize the candidate service composition, in the set of candidate services, service composition generator can find the optimal service composition solution.In the process of service composition optimization based on semantic web service composition trust model, an ant colony algorithm is proposed and implemented to adapt to the change of the path weight. Before the global pheromone update of the local optimal solution, the adaptive ant colony algorithm will get the weight of the local optimal solution on the path. The change of the weight match a certain standard, the new algorithm will update the weight to adapt the environment. And the new algorithm will dynamic adjust pheromone evaporation factor to avoid the phenomenon of premature convergence of the algorithm. In final, designs an example of the service composition of a travel itinerary. Through the experiments, prove that the service portfolio selection and optimization method based on the semantic web service composition trust degree model can effectively avoid the false information of the non-trusted component services, and guarantee the final solution can meet user needs very well.