Parameters setting and analysis for ant colony optimization algorithm

摘要

This paper introduces swarm intelligence, ant colony algorithm principle and their advantages and disadvantages. A large number of experiments based on the parameters choice of m, α, β, ρ, Q were studied, the law of their choice was concluded, and "four steps" was put forwards on the basis of "three steps" in the past, which was an effective way for ant colony algorithm to select the optimal combination of parameters, and then improved ant colony algorithm was analyzed, while an experimental approach for these types of algorithms including optimal retention policy ant system, largest - smallest ant system, ant-based sort system and best-worst ant system on performance of TSP were compared with and analyzed, and the results of the performance were ranked, and when the TSP problems based on other properties with the same conditions (including number of iterations, the iteration time, etc.) have optimal results.