Decentralized Semantic Coordination Through Belief Propagation

This paper proposes a generic decentralized method for interconnected entities to compute globally coherent sets of mappings with respect to the semantics of ontological specifications and their subjective mapping preferences. This problem is formed as an optimization problem between interdependent agents. Globally coherent sets of mappings are computed by means of a distributed extension of the maxplus algorithm, taking also into account the feedback entities receive on their subjective mappings from distant others. Experimental results from a large number of networks of varying complexity show the strengths of the proposed approach and point to further work.