Minimal Model Generation with Factorization and Constrained Search

概要

論文の詳細を見る
Herbrand models are useful in several areas of computer science. In most cases, however, the native model generation algorithms are not suitable because they return non-minimal models which are often redundant. This paper describes an approach for generating minimal Herbrand models. The approach is based on model generation with factorization which is shown to be complete for generating minimal Herbrand models. Factorization can prune more search spaces than complement splitting does. Complement splitting is widely used for eliminating redundant search. In order to generate only minimal Herbrand models, constrained search strategy is applied. We give its two implementations as short Prolog programs. We evaluated their performance by proving some typical problems.
一般社団法人情報処理学会の論文
2003-04-15