Ubiquitous Collaborative Multi-terminal Network Of Trust Mechanism

At present, many countries and cities are promoting research and development of the "Ubiquitous City" and "Smart City", hoping to have an opportunity to drive the whole information industry, and even the development of economic construction. But to truly achieve Ubiquitous Network, the core problem is how heterogeneous terminals communicate credibly between each other. We will achieve communication and "Ubiquitous" between varieties of heterogeneous terminals only if more reliable interaction through heterogeneous terminals could be made.However, in existing trust models the recommending services often provide a single trust value (computed by them during their interaction with the unknown entity in question) as a recommendation. However, a single trust value recommended by a recommender represents its subjective opinion about the unknown entity and cannot depict the real trust level very well under certain circumstances. To address this issue, firstly a fuzzy based credibility evaluation method for indirect trust computation is proposed. The calculation method cannot just incorporate mechanism to determine the weight each valid recommendation should carry in aggregation process but also distinguish between honest and dishonest recommendations in the meantime.Secondly,the paper puts present a Comprehensive Trust Model based on Reputation and Fuzzy subsystems (CFMBRF), indirect recommending trust calculation method of the model adopts the "an fuzzy based credibility evaluation method for indirect trust computation", then on the basis of the model based on reputation, import the fuzzy inference subsystems, it is able to handle subjective concept such as " importance of a interaction","the decisions in the uncertainty region " and "setting the result of interaction ", can humanistic make decision about uncertain problem.Finally, the paper verify effectiveness of the "a fuzzy based credibility evaluation method for indirect trust computation". The results show that the calculation method cannot just incorporate mechanism to determine the weight each valid recommendation should carry in aggregation process but also distinguish between honest and dishonest recommendations in the meantime. On the other hand, the paper simulates the network having no-fuzzy subsystem and network which have the fuzzy subsystem, makes comparison between the two networks, tests overhead and stability of the CFMBRF. It can be seen that the network using the fuzzy subsystem greatly increased the validity of the interaction, decreased the number of query for the reputation vector,proved the effectiveness of the CFMBRF model.