ASnode-SCP: A new version of the Ant System for the Set Covering Problem

Ricardo M. A. SilvaGeber L. Ramalho

Ant systems are a novel heuristic method for solving hard combinatorial optimization problems. This paper investigates the feasibility of ant systems to a category of facility location problem: the set-covering problem (SCP). This work shows that the straightforward adaptation of the original ant system model to the set-covering problem (AS-SCP) depends strongly on a domain-dependent heuristic to achieve good results. So, a novel model is proposed, called ASnode-SCP, which does not use any heuristic information concerning the problem. This new model is compared to AS-SCP and to stochastic greedy search to demonstrate experimentally that it can rapidly provide good, if not optimal, solution. Clique no link abaixo para buscar o texto completo deste trabalho na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: