Validation of knowledge-based systems by means of stochastic search
Laure Brisoux Devendeville, Éric Grégoire, L. Sais · 2002
The use of stochastic search to validate first order knowledge based systems is investigated. It is well known that such techniques can prove efficient in showing that consistency constraints do hold, at least in the propositional case. Powerful heuristics about the trace of stochastic search allow proofs of inconsistency to be obtained as well. But, how stochastic search can be applied to first order knowledge bases without giving rise to a combinatorial space explosion remains an open issue. A partial instantiation schema is proposed in the context of the incremental consistency/inconsistency problem. It allows forms of depth limited consistency and inconsistency to be handled in an effective manner showing promising paths for the development of new efficient consistency checking techniques for first order knowledge bases.