A clustering-based adaptive parameter control method for continuous ant colony optimization

Ant colony optimization (ACO) has been widely and successfully applied to NP-hard combinatorial optimization problems for its strong searching ability and robustness. Recently, several extended ACO algorithms have also been proposed to deal with continuous optimization problems. However, the ACO algorithms always have slow convergence speed and encounter premature convergence in engineering applications. This paper proposes a novel adaptive parameter control method for continuous ACO algorithms. Clustering analysis is used to judge the optimization state of the algorithm and the flexible adjustment of the parameters is based on these optimization states during the training process. As an example, the adaptive control method is used to improve the performance of the continuous orthogonal ant colony (COAC). Experimental results demonstrate that the clustering-based adaptive parameters control scheme contributes to both faster convergence speed and higher solution accuracy. The proposed adaptive control method has great practical value and bright prospect.