ASTREA: Answer sets for a trusted reasoning environment for agents

In recent years, numerous papers have shown the power and flexibility of answer set programming (ASP) in the modeling of intelligent agents. This is not surprising since ASP was developed specifically for non-monotonic reasoning - including common-sense reasoning required in agent modeling. When dealing with multiple agents exchanging information, a common problem is dealing with conflicting information. As with humans, our intelligent agents may trust information from some agents more and than from others. In this paper, we present ASTREA, a methodology and framework for modeling multi-agent systems with trust. Starting from agents written in standard , we model the agent's knowledge, beliefs, reasoning capabilities and trust in other agents together with a conflict resolution strategy in CR-Prolog. The system is then able to advise the agent what information to take into account and what to discard.