An ant colony algorithm aimed at dynamic continuous optimization

The introduction of the concept of swarm intelligence into ant colony optimization (ACO) algorithms has shown the rich possibilities of self-organization when dealing with difficult optimization. Indeed, the inherent flexibility and efficiency of ACO algorithms proved to be advantageous for difficult dynamic discrete problems, e.g. routing in telecommunication networks. Moreover, we believe that ant colony algorithms can be efficient for both continuous dynamic problems and discrete ones. In order to exploit the features of these swarm intelligence algorithms for continuous dynamic optimization, we introduce an hybrid population-based ant colony algorithm. Considering the way ants communicate, we propose a “heterarchical” algorithm, called “Dynamic Hybrid Continuous Interacting Ant Colony” (DHCIAC), based on the hybridization of an “interacting ant colony” with a Nelder-Mead algorithm. Being confronted with the lack of benchmark functions for dynamic optimization in the literature, we have elaborated a complete set of various continuous dynamic problems. The efficiency of the proposed DHCIAC algorithm is then demonstrated through numerous tests, conducted involving that new benchmark platform.