A new hybrid heuristic approach for solving large traveling salesman problem

This paper presents a new metaheuristic approach called ACOMAC algorithm for solving the traveling salesman problem (TSP). We introduce multiple ant clans' concept from parallel genetic algorithm to search solution space utilizing various islands to avoid local minima and thus can yield global minimum for solving the traveling salesman problem. Moreover, we present two approaches named the multiple nearest neighbor (NN) and the dual nearest neighbor (DNN) to ACOMAC to enhance large TSPs. To validate the proposed methods, numerous simulations were conducted to compare ACOMAC and Dorigo's ACS with and without the addition of the multiple nearest neighbor (NN) method or the dual nearest neighbor (DNN) approach, using a range of TSP benchmark problems. According to the results of the simulation, adding the NN or DNN approach to ACOMAC or ACS, as initial solutions, also significantly enhances the performance of ACOMAC and ACS in solving the traveling salesman problem. Meanwhile, using ACOMAC + DNN with TSP can yield better solutions than the other stated approaches. Additionally, ACOMAC or ACOMAC + NN, utilizing five ant clans with a total of 20 ants, is verified to yield better solutions. Furthermore, ACOMAC with a local weighting (w) set to 0.6 can yield better solutions in terms of length.