大洪水算法是通过模拟洪水上涨过程来进行全局寻优的启发式算法,r-opt算法是一类常用的路径改进算法.本文针对旅行商问题,提出一种将二者有机融合的改进大洪水算法,可用于快速求解大规模和超大规模的TSP问题.算法在Delphi7环境下编程实现,经过大量TSPLIB中的数据实例进行测试和验证,求解结果与已公布的最好结果误差基本都在1%以下,为困难的大规模旅行商问题提供了新的求解手段.
Great deluge algorithm is one" of heuristics by simulating the process of flood rising to search global optimization, r-Opt algorithm is usually applied in path optimization. For the travelling salesman problem, this paper gives a modified great deluge algo- rithm which mainly combines great deluge algorithm with r-opt algorithm, and it can be used to solve the large-scale and super large- scale travelling salesman problem. The algorithm is programmed in Delphi 7, and is tested through series of standard instances from TSPLIB. The errors between the algorithm results and the best results published in TSPLIB are almost below 1%. The algorithm is proved to be a kind of new method to solve the difficult large-scale travelling salesman problem.