A new design and implementation of progol by bottom-up computation

Abstract

This paper describes a parallel version of Progol based on MGTP which is a theorem prover employing bottom-up inference suitable for parallel implementation. Hypothesis formation in Progol, which is performed by top-down computation with Prolog in the sequential implementation, will be performed more efficiently by bottom-up computation with MGTP in the new implementation. For the Progol's generalto-specific search for hypotheses through the subsumption lattice, we developed a new way of calculating a heuristic function for the A*-like algorithm, which was also implemented with MGTP. Since MGTP already has very efficient parallel implementations on parallel inference machines, an efficient implementation of parallel-Progol will readily be realized as well.