SOLAR: An automated deduction system for consequence finding

Issue title: Practical Aspects of Automated Reasoning

Article type: Research Article

Authors: Nabeshima, Hidetomo | Iwanuma, Koji | Inoue, Katsumi | Ray, Oliver

Affiliations: University of Yamanashi, Yamanashi, Japan. E-mails: {nabesima, iwanuma}@yamanashi.ac.jp | National Institute of Informatics, Chiyoda-ku, Japan. E-mail: ki@nii.ac.jp | University of Bristol, Bristol, UK. E-mail: oray@cs.bris.ac.uk

Abstract: SOLAR (SOL for Advanced Reasoning) is a first-order clausal consequence finding system based on the SOL (Skip Ordered Linear) tableau calculus. The ability to find non-trivial consequences of an axiom set is useful in many applications of Artificial Intelligence such as theorem proving, query answering and nonmonotonic reasoning. SOL is a connection tableau calculus which is complete for finding the non-subsumed consequences of a clausal theory. SOLAR is an efficient implementation of SOL that employs several methods to prune away redundant branches of the search space. This paper introduces some of the key pruning and control strategies implemented in SOLAR and demonstrates their effectiveness on a collection of benchmark problems.

Keywords: Consequence finding, theorem proving, SOL calculus, connection tableaux, first-order logic

DOI: 10.3233/AIC-2010-0465

Journal: AI Communications, vol. 23, no. 2-3, pp. 183-203, 2010