A Parallel Implementation of Ant Colony Optimization

Abstract

Ant Colony Optimization is a relatively new class of meta-heuristic search techniques for optimization problems. As it is a population-based technique that examines numerous solution options at each step of the algorithm, there are a variety of parallelization opportunities. In this paper, several parallel decomposition strategies are examined. These techniques are applied to a specific problem, namely the travelling salesman problem, with encouraging speedup and efficiency results.