An automata view to goal-directed methods

Abstract

Consequence-based and automata-based algorithms encompass two families of approaches that have been thoroughly studied as reasoning methods for many logical formalisms. While automata are useful for finding tight complexity bounds, consequence-based algorithms are typically simpler to describe, implement, and optimize. In this paper, we show that consequence-based reasoning can be reduced to the emptiness test of an appropriately built automaton. Thanks to this reduction, one can focus on developing efficient consequence-based algorithms, obtaining complexity bounds and other benefits of automata methods for free.