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.