Multi-Engine Answer Set Programming

VCLA will be hosting a talk by Marco Maratea on October 27, 2014.

ABSTRACT

It is well-known that, on empirically hard problems, there is rarely a “global” best algorithm. Instead, different algorithms perform well on different problem domains/instances. This fact can be taken as an advantage, by exploiting machine learning techniques. Approaches based on the "multi-engine" recipe exploit some of such techniques. In this talk I will present our work on the design, implementation and experimental evaluation of a multi-engine solver for Answer Set Programming (ASP), called ME-ASP. I will start by describing the approach on ground ASP programs; then, I ll move to a more recent work where a similar recipe is applied to non-ground programs.