Using branching time logic to optimize an extended class of Datalog queries

Petros Potikas, Manolis Gergatsoulis, Panos Rondogiannis

Abstract: We propose an extension of the branching-time transformation which can handle a significantly broader class of Datalog programs. The initial transformation could only be applied to Chain Datalog, a useful but restricted class of programs. In this paper we demonstrate that the branching-time transformation can be extended to handle all Datalog programs that do not allow multiple consumptions of variables in clauses. We demonstrate the correctness of the new transformation and provide certain optimizations that further improve the programs obtained by the transformation.
Keywords: Deductive Databases, Temporal Logic Programming.