Optimization based on multi-type ants for the Traveling Salesman Problem

摘要

This paper proposes an algorithm based on multi-type ants (so-called Multi-type Ant Colony System, M-ACS) for solving the Traveling Salesman Problem (TSP). Multiple ant types cooperate (through pheromone exchange) and compete (by mean of repulsion mechanism) among them in order to increase the efficacy in the search process. The obtained experimental results has been compared against benchmark results from OR literature. In particular for large scale TSP, the algorithmic proposal M-ACS shows competitive results regarding to the efficacy.