Parallel ACS for weighted MAX-SAT

Abstract

The ant colony system or ACS is an important approach of the Ant Colony Optimization meta-heuristic. In this work we aim to study the efficiency of ACS in solving the weighted max-sat problem. This will require an adaptation of all the rules of this approach to the elements which characterize our problem. All the ACO algorithms contain a low dependence level, this feature makes them beneficent to parallelize. Thereby, we will also study various ways to parallelize the ACS algorithm proposed in the first step of this work and discuss their impacts and differences. © Springer-Verlag Berlin Heidelberg 2003.