Although ant colony optimization (ACO) has successfully been applied to a wide range of optimization problems, its high time- and space-complexity prevent it to be applied to the large-scale instances. Furthermore, local search, used in ACO to increase its performance, is applied without using heuristic information stored in pheromone values. To overcome these problems, this paper proposes new strategies including effective representation and heuristics, which speed up ACO and enable it to be applied to large-scale instances. Results show that in performed experiments, proposed ACO has better performance than other versions in terms of accuracy and speed.
