Author(s): ZHANG Chi, TU Li, WANG Jiayang, School of Information Science and Engineering, Central South University, School of Information Science and Engineering, Hunan City University
Pages: 2944-2949
Year: 2015 Issue:  8
Journal: Journal of Central South University of Technology
Keyword:  ant colony algorithm(ACA); roulette; pheromones; differential evolution; convulsions;
Abstract: Considering that there exists precocious and stagnation behavior phenomenon about the traditional ant algorithm, a new ant colony algorithm(NACA) was proposed, i.e. the transition rule, convulsions rule of global pheromones and the mixing adjustment rule of local pheromones were mixed. The results show that on one hand, NACA enhances genetic algorithms population diversity and at the same time the roulette operator is used in translative rules, whch is beneficial to improving the convergence speed. On the other hand, the individual differential information is used in the updated rule of local pheromones, which is good for searching the global solution. Finally, the cataclysm operator avoids falling into local optimum, and leads to reach the global optimum. NACA has stronger ability to search the global optimal solution and better stability and astringency. It also provides a new idea for solving other combinational optimization problems.