Research On The Algorithm Of The Web Service Composition Based On Trust

In recent years, the research of Web service is becoming more and more mature,however, there is a lack of research on the service reliability in the implementation scheme of Web service, and there is no mature standard to ensure the reliability of the service. Trust issues and the study of trust model, is the main method of solving reliability problems in various fields, its application in Web services and service portfolio as an alternative basis for one service, choosing a more reliable service, and thus effectively improve task execution completion rate.This thesis is aimed at the problem of Web service trust in cloud computing, first in order to complete the assessment of the service entity of trust, introduced Bayesian formula and Bayesian model to build trust model, that according to the interaction between these service entities, a direct trust value and indirect trust value assigned to each entity, and the value distribution of weight factor for the two kinds of trust, thus integrated service entities trust value is calculated by the weighted sum method, at the same time, to prevent malicious entities impact on the real value of trust services,introduce reward and punishment factor, in order to send the trust evaluation feedback to the service provider well, and finally achieve the goal of updating the direct trust value.Next, the QoS attribute is used as an important parameter in the service composition model, and combine the trust and QoS attributes, select the cuckoo algorithm for service composition by comparing various selection optimization algorithm, it worked out the problem that the general particle swarm algorithm,Skyline and other methods in respect of the multi-object function versatility and superiority is not strong.Finally, select the QWS dataset as a basic research, build a service composition model. Simulation experiments are carried out on the basis of the model, making trust as a parameter and set it to the QoS parameters,then using the cuckoo algorithm design and analysis phase of the service portfolio. In order to verify the feasibility and effectiveness of the service composition model created in this thesis, compared with the service composition model based on particle swarm optimization, by analyzing data on the experimental results proved that cuckoo algorithm is become moreexcellent in solving the problem of the multiple objective function optimization, while demonstrated the validity of service selection which is realized with trust model.