Modified Ant Colony Optimization with Updating Pheromone by Leader and Re-Initialization Pheromone for Travelling Salesman Problem

Abstract

Ant Colony Optimization (ACO) algorithm is a stochastic algorithm. It is used for solving combinational optimization problem. The ant colony walks along density of pheromone from ant's nest to feeding sources. It leads to create shortest path from ant's nest to feeding sources. Normally, ACO encounters the problem of trapping in local optimum. To improve solutions, 2-Opt algorithm is applied with ACO. However, 2-Opt algorithm cannot solve trapping in local optimum of ACO and cannot improve searching performance of ACO. This paper proposed improving ACO algorithm by the results from searching of 2-Opt algorithm are applied with pheromone of ants. Moreover, when ant colony occur trapping in local optimum, the pheromone of ants is re-initialized to solve trapping in local optimum problem. The proposed technique is tested on twenty-three maps from the Traveling Salesman Problem Library (TSPLIB) and gives more satisfied search results in comparison with ACOs.