Trust evaluation model based on multiple service attributes

when the most trusted models evaluate level of trust about service entities at present, the credibility of the service attributes whether the two sides trading result is credible is rather ambiguous, which makes the evaluation model lack of persuasion, and not very good depict the complexity and uncertainty of trust relationship between the two sides. In view of the above questions, learning trust relationship from human society, according to the service of the node itself, abstracting representative multiple service attributes from the trusted relationship of the node service, service requester comprehensive evaluates multiple service attributes provided by service provider based on personal interest, and combines with the trust value, decide whether to trade finally; after the transaction, service request calculates QoS(quality of service) difference degree according to actual QoS and service provider'own QoS claimed to judge the credibility of service provider, then gives corresponding rewards and punishment and trust update. Simulation results show that the model can evaluative accurately the node's trust degree and effectively control of malicious nodes attack, when the malicious nodes have a higher proportion in the system, it can still keep high success rate. © 2011 ACADEMY PUBLISHER.