Improved Bat Algorithm for Solving Traveling Salesman Problems
Hongxia Zhao · 2023
To solve traveling salesman problems more efficiently, an improved parallel bat algorithm (IPBA) is proposed in this research. The main improvement measures are as follows. First, to be suitable for solving discrete problems, the relevant operations of the algorithm are discretized. Then, a chaotic initialization based on the immune concentration concept is presented to obtain a more competitive initial population. In addition, decreasing inertia weight and stagnate variation strategy are adopted to balance the global and local search of the proposed algorithm. Especially, parallel strategy is introduced to divide the whole population into two subpopulations of exploration and exploitation, which adopts different inertia weights and stagnate variation strategies to evolve. Regular information exchange between subpopulations is used for the advantages of parallel strategy and improves the overall performance of the proposed algorithm as well. At last, the proposed IPBA is adopted to solve typical traveling salesman problems with different scales, and the comparison and analysis of the optimal results verify the feasibility and effectiveness of the proposed algorithm.