INVESTIGATION OF CONVERGENCE PROPERTIES OF ANT COLONY BASED ALGORITHMS

Abstract

The paper presents the Ant Colony Optimization (ACO) algorithms, and investigates their convergence properties. Our investigation is limited to the following ACO algorithms: Ant System, MAX-MIN Ant System, Elitist Strategy of Ant System and Rank Based Version of Ant System. In the literature can be seen, that a novel of optimization problems are solved with ACO algorithms. In our investigation, we limited the Multi-Depot Vehicle Routing Problem with Time Windows (MDVRPTW), which is an NP-hard discrete optimization problem. The paper presents a literature review in connection with ACO algorithms, and Vehicle Routing Problem (VRP). After that, the mathematical model is written. The paper also presents the implemented algorithms, and the test results based on benchmark data.