Object-Guided Ant Colony Optimization Algorithm with Enhanced Memory for Traveling Salesman Problem

In this study, we presents an object-guided ACO algorithm which is consisted of ants with enhanced memory. In the process of solution construction, each ant stores a complete solution in its enhanced memory. Each time ant selects a solution component probabilistically, it will calculate the difference between current solution and the new solution after adding the selected component and then Metropolis accepting rule, which has been used in simulated annealing algorithm successfully, is used to decide whether to accept the component or discard it. The simulation results, which were carried on benchmark traveling salesman problems, show that the improvement of individual intelligence can improve the performance of ACO algorithm remarkably.