Parameter Analysis for a Novel Ant Colony Optimization Algorithm

Abstract

Ant colony optimization (ACO) is a class of stochastic search procedures working in the space of the solutions which has been applied to several NP-hard combinatorial optimization problems. Two-stage updating pheromone for invariant ant colony optimization algorithm (TSIACO) as one of novel ACO algorithm has been proved that it is a strong invariant ACO algorithm. Experimental results based on traveling salesman problems (TSP) show its feasibility compared to max-min ant system (MMAS). However, how the parameters affecting the performance of TSIACO has not been studied. In this paper, the framework of constructing invariant ACO algorithm based on TSIACO is proposed firstly. Then, we use the experimental analysis method to study the action of parameters based on TSP. Lastly, we compare the performance of TSIACO with other novel ACO algorithms to show the effectiveness of TSIACO.