On Computing Non-hypocritical Consensuses in Standard Logic
Éric Grégoire, Jean-Marie Lagniez, Du Zhang · 2016
In this paper, we improve a computational concept of logic-based consensus among agents that has been introduced recently. A consensus is defined as a non-contradictory fragment of the information conveyed by various agents such that this fragment does not logically conflict with any agent. In this study, we introduce an additional requirement to the concept of consensus to avoid some paradoxical consensuses. Intuitively, we provide a mechanism that prevents some undesirable counterfactuals from being generated in the consensus formation process when an intuitive ad-hoc encoding of "If then" rules in standard logic is used.