A genetic algorithm for traveling salesman problems
Huai‐Kuang Tsai, Jinn‐Moon Yang, Cheng‐Yan Kao · 2001
This paper proposes an evolutionary approach for the traveling salesman problem. The proposed approach consists of global and local strategies by incorporating the family competition into edge assembly crossover and near 2-opt mutation. The method is applied to six well-known problems, including the eil101, lin318, pcb442, att532, rat575, and u724. The experimental results indicate that the proposed approach performs robustly and is very competitive with the other approaches surveyed in this paper. As our approach, although somewhat slower, executes 50 independent runs for the att532 problem, it is able to find the optimum solution in 23 runs and the average value of solution quality is 27691.3 To the best of our knowledge, the solution quality is the best in our surveys for this problem.