A Trust Evaluation Model Of Web Services Based On Requirement Context And Recommendation

Along with the development of the distributed application, the distributed model based on web service has become a kind of typical Internet application form under the network environment. But because of the openness and deceptive characteristics of web service environment, many researches on various trust models of estimating the trust between trading entity appeared in order to reduce transaction risks and ensure the safety of trading environment. But because there is no unified concept of trust and there are all kinds of malicious entities existencing, how to assess the trust of two entities more accuratly and reliably and how to test and cripple various malicious entities' malicious behaviors has become a key issue.In this paper, on the basis of the existing research on trust mechanism and through the study of web service, we proposed a trust evaluation model of web service which based on requirement context and recommendation, the paper explored from the following aspects:First, according to expanding the existing trust model based on feedback,we put forward the concept of requirement context,and use it to mark the trust relationship between two entities.Then we presented a DTCA algorithm according to requirement contexts'similarity and constraint relationship to calculate direct trust, in order to improve the accuracy of assessment of direct trust.Secondly, put forward the concept of recommendation honesty and recommendation dishonesty,and according to the two aspects using fuzzy reasoning method to calculate recommenders'credibility.Through the method,we can distinguish the ability of different recommenders'and weaken the influence of various malicious entities.The model considering the relationship between requestor and recommenders,requirement context and preference similarity factors in the calculation of indirect trust,is aim to enhance the accuracy and reliability of indirect trust assessment.Third, Studied the evolution mechanism of trust from time attenuation factor, trading scale factor and punish factor to revised the trust model, and provides a competition mechanism to update friends and strangers relationship. Through the improved method we aim to enhance the reliability of the models and incentive entities'goodwill behavior.Fourth,A two-way trust decisions making method takes into account the service requesters'and providers'trust requirements in order to reduce the provider's trading risk and encourage the requester to provide honest feedback initiatively.Finally, through simulation experiments we verify that the WEBTRM model proposed in this paper is feasibility and effectiveness.