Quasi-merging and Pure-arbitration on Information for the Family of Adaptive Logics ADM∗

The present paper introduces two new information merging protocols for the family of adaptive logics ADM, for which majority merging has been defined in previous work. The new adaptive operators reflect the negotiation processes of quasi-merging and pure-arbitration known from the Integrity Constraints framework. The Adaptive Variant Counting selection provides results equivalent to the GMax family of merging operators: it selects a collective model for a multi-set of belief bases established on the number of disagreements verified by the various models. The Adaptive Minimax Counting selection is a quasi-merging operator which applies a minimax function and it obtains a larger spectrum of possibilities than the previous selection: it simulates the behaviour of the Max family of operators from the Integrity Constraints framework, but it avoids some of its counterintuitive results.