Distributed Query Answering in Peer-to-peer Reasoning Systems Technical Report

Interest in distributed reasoning is gaining momentum due to the emergence of the Semantic Web. In this work, we address the problem of meaningfully answering queries to distributed propositional reasoners which may be mutually inconsistent. We define a peer-to-peer query answering framework in which peers may be prioritized. Our formal framework includes distributed entailment relations for both consistent and inconsistent systems. We develop decentralized algorithms for computing answers to a restricted class of queries according to distributed entailment and prove their soundness and completeness. A heuristic for more efficiently computing query answers from prioritized, inconsistent systems is also provided and its effectiveness demonstrated empirically. We furthermore investigate global closedworld reasoning in our framework and show how our techniques can be used to achieve global generalized closed-world query answering from local closed-world reasoners in a restricted class of distributed peer-to-peer query answering systems.