Trust Estimation in autonomic networks: a message passing approach

The ability to maintain belief relationship among entities in autonomic networks is considered a major challenge. In this work we tackle the problem by casting it into the framework of Estimation Theory as an inference problem on a Markov Random Field. A fully distributed algorithm based on message passing techniques is then proposed, where messages are not considered as abstract intermediate results of a computation, but as real messages exchanged by the nodes in the network. With this case study we therefore demonstrate that Markov Random Field theory used in combination with Message Passing algorithms constitutes a powerful theoretical framework for the development of algorithms for information distribution and fusion.