Adaptive arbitration by variant counting on commutative bases with weights

In this paper a new logical arbitration protocol for fusion of inconsistent information is designed. It defines a selection of models of a premise set in a multi-modal logic that uses the standard format of adaptive logics. The selected models are obtained by a counting procedure on the derivable data conflicting among the various sources. Peculiar of this approach is the definition of weights for commutative bases, in terms of the distinction between partially and fully supported information. The results obtained are compared to standard arbitration protocols and they extend previous work on the adaptive majority protocol.