Parallel answer set programming

Abstract

Answer Set Programming (ASP) has become, in recent years, the paradigm of choice for the logic programming community and for a wide variety of application domains. Thanks to its declarative nature, ASP offers excellent opportunities for performance improvements through transparent exploitation of parallelism. This Chapter provides a survey on the main techniques and approaches in the literature to enable exploitation of parallelism in the execution of Answer Set Programming solvers. The survey explores the approaches along two orthogonal dimensions. The first dimension considers the different levels of complexity and features of the underlying language, ranging from propositional Datalog/definite Horn clauses to full ASP. The second dimension, instead, explores the different levels of granularity of exploitation of parallelism, ranging from fine grain parallelism, exploited using general-purpose graphical processing units, to very large grain parallelism exploited on distributed platforms.