Improved Discrete Wolf Pack Algorithm for Solving Travelling Salesman Problmes

To solve the travelling salesman problems (TSP) more satisfactorily, an improved discrete wolf pack algorithm (IDWPA) is proposed based on original wolf pack algorithm (WPA). Firstly, we apply the adaptive mechanism into three intelligent behaviors of the wolf pack within WPA, and discretize them respectively to be suitable for solving TSP problems. This mechanism can improve optimization efficiency by balancing exploration and exploitation at different stages of algorithm execution. Then, in order to further improve the convergence speed, renewal strategy based on hunger values is introduced instead of the stronger-survive renewing rule for the wolf pack. Finally, several TSP problems on different scales selected from TSPLIB library are resolved by proposed IDWAP, discrete WAP (DWAP) and other typical optimization algorithms. Simulation results show that IDWAP is more superior to others. We can obtain shorter optimal paths by proposed IDWAP with high probability. And the feasibility and effectiveness of IDWAP are verified as well.