A novel Max-Min ant system algorithm for traveling salesman problem

A novel max-min ant system algorithm is proposed for the problem of setting pheromone trail value which is caused by uncertainty of the objective function value. The upper and lower bounds of pheromone are determined according to the range of objective function value by a random sampling. And the update quantity of pheromone is determined. As result of this process is not using the objective function value, so they are independent of it. Then, pheromone is updated by two phases to keep an appropriate balance of the exploration and the exploitation. In the early phase of algorithm, the first l iterative optimal solutions are employed to enhance exploration capability. Then, in the later phase of algorithm, iteration-best optimal solution is used to update pheromone tail to accelerate the convergence rate. An example of traveling salesman problem is given, which is simulated by proposed algorithm, max-min ant system, and other improved ant algorithms. Simulation results show the effectiveness of the proposed algorithm.