On Evaluating Decision Procedures for Modal Logics

This paper investigates the evaluation method of decision procedures for multi-modal logic proposed by Giunchiglia and Sebastiani as an adaptation from the evaluation method of Mitchell et al.\ of decision procedures for propositional logic. We compare three different theorem proving approaches, namely, the Davis-Putnam-based procedure K\textsc{sat}, the tableaux-based system $\mathcal{KRIS} and a translation approach combined with first-order resolution. Our results do not support the claims of Giunchiglia and Sebastiani concerning the computational superiority of K\textsc{sat} over $\mathcal{KRIS}$, and an easy-hard-easy pattern for randomly generated modal formulae