Inverse Problemin DSmT and Its Applications in Trust Management

DSmT (Dezert-Smarandache Theory) is actually a natural extension of DST (Dempster-Shafer Evidence Theory) and Bayes Probability Theory, which can handle both uncertainty and contradiction efficiently. We present the inverse problem in DSmT and research on the existence of its unique exact solution. In particular, we adopt PSO (Particle Swarm Optimization) arithmetic to find an approximate solution if the inverse problem dose not have a unique exact solution. Trust management has become a key technology and hot topic in open computational systems in recent years. We apply DSmT to the trust management field as a new way of trust representation. Moreover, the simulation suggests that the difference between the second-hand evidential trust evaluation and the first-hand evidential trust evaluation can be greatly reduced by the modification based on the inverse problem framework in DSmT.