Ant colony optimization with self-adaptive parameter adjustment

This paper first introduced the rough searching and subtle searching process of ant colony optimization (ACO) in solving multi-dimensional and constrained functions, and then analyzed the key parameters affecting the optimization performance of ACO. A conclusion was drawn which required that the parameter α and β should be adjusted in descending order and ρ in ascending order. Next three fuzzy dynamic regulators of α, β and ρ were built respectively, and their corresponding parameter adjustment procedures, the executed strategies and the control procedures were also given. Finally the performance of the proposed ACO with adaptive parameter adjustment was verified in the case of a crane girder´s optimization. Comparison results show that the self-adaptive adjusting ACO has many advantages such as higher solution precision, higher optimization efficiency and less number of ant colonies over conventional ACO. This method is a promising optimization algorithm for solving complex optimization problems.