A novel recommendation trust revision algorithm for autonomous networks

The performance of autonomous networks depends on collaboration among distributed entities. To enhance security in autonomous networks, it is important to correctly revise the trustworthiness of recommendations of entities since trust management is the main foundation of collaboration while it faces many new attacks. In this paper, new attacks against trust evaluation are identified and relationships between these attacks are analyzed. Then we present a mathematical method for recommendation trust revision according to the deviation of recommendation. This method is based on Bayesian decision-making theory. The distribution of random value deviation of recommendation is described in the beta distribution and recommendation trust will be revision according to the principle of minimum loss function. Our study shows this method is helpful to reduce the impact of some new threats to trust management.