Optimal Routing in Traveling Salesman Problem using Artificial Bee Colony and Simulated Annealing

Traveling salesman problem (TSP) is a popular pathselection problem, which is a sub-problem of many applicationdomains such as transportation, network communication andvehicle routing. TSP belongs to the class of NP-hard problems.Among many approaches which have been proposed for TSP,evolutionary and swarm intelligence algorithms efficientlyapplied to solve it, and attempt to avoid trapping in localminima. In this paper, a hybrid algorithm based on ArtificialBee Colony (ABC) and Simulated Annealing (SA) is proposedfor TSP. The proposed algorithm, named ABCSA, is dividedinto two phases. The first phase searches globally the searchspace via population-based ABC algorithm. The second phaseapplies local-based SA algorithm in order to improve the finalsolution of ABC. Instead of a random solution, the final bestsolution gathered by ABC is considered as the initial solutionof SA. To demonstrate the effectiveness and efficiency of theproposed algorithm, some benchmark problems from TSPLIBwere tested and compared with the techniques of GA, PSO, SAand ABC. Results show that our algorithm achieved shorterdistances in all cases with fewer generations and running time.

Keywords:

traveling salesman problem , routing , neighborhood search , artificial bee colony , simulated annealing