Parameters Analysis for Basic Ant Colony Optimization Algorithm in TSP

Abstract

In order to effectively address the lack of basic ant colony algorithm in terms of parameters, we use four-step method instead of the popular three-step, based on a large number of experiments of the parameters setting, this paper summed up an effective selection method for m, α, β, ρ and Q parameters to select the best combination of parameters. Applying the improved ant colony algorithms including optimal retention policy ant system, max-min ant system, ant-based sorting systems and best-worst ant system, performance comparison analysis was conducted with the same TSP problems, and experiments proved that the proposed method of parameter combinations greatly improves the speed of convergence.