A COMPARATIVE STUDY OF SOME METAHEURISTIC ALGORITHMS FOR SOLVING TRAVELLING SALESMAN PROBLEM

Application of the following three metaheuristic algorithms to Travelling Salesman Problem (TSP) were explored: Simulated Annealing (SA), Tabu Search (TS), and Ant Colony System (ACS). The performance of software implementation of these approaches was experimentally studied using four test instances of TSP with known optimal solutions. As a result of the experiment, features of solutions found in a set of trials by each algorithm in the same fixed amount of time were assessed. The implementation details of the algorithms and the parameters used for the experiment were thoroughly described. Optimization process was visualized for every algorithm. Conclusions were made regarding the effectiveness of studied algorithms for the different sizes of problem instances used in the experiment.