Finding Maximum Cliques with Distributed Ants

In this paper we describe an ant system algorithm (ASMC) for the problem of finding the maximum clique in a given graph. In the algorithm each ant has only local knowledge of the graph. Working together the ants induce a candidate set of vertices from which a clique can be constructed. The algorithm was designed so that it can be easily implemented in a distributed system. One such implementation is also described in the paper. For 22 of the 30 graphs tested ASMC found the optimal solution. For the remaining graphs ASMC produced solutions that are within 16% of the optimal, with most being within 8% of the optimal. The performance of ASMC is comparable to existing algorithms.

• Type: Conference
• Year: 2004
• Where: GECCO
• Authors: Thang Nguyen Bui, Joseph R. Rizzo
• Keywords: Ant System Algorithm, GECCO 2004, Graphs Tested Asmc, Maximum Clique