Parallel Ant System for the Set Covering Problem

The Set Covering Problem (SCP) represents an important class of NP-hard combinatorial optimization problems. Actually, the exact algorithms, such as branch and bound, don't find optimal solutions within a reasonable amount of time, except for small problems. Recently, the Ant systems (AS) were proposed for solving several combinatorial optimization problems. This paper presents a hybrid approach based on Ant Systems combined with a local search for solving the SCP. To validate the implemented approach, many tests have been realized on known benchmarks, and, an empirical adjustment of its parameters has been realized. To improve the performance of this algorithm, two parallel implementations are proposed.