一种改进的自组织映射算法求解旅行商问题

摘要

目前，没有求解旅行商问题的非常有效的方法。提出了一种求解该问题的LNSOM算法，在自组织映射算法的基础上，改进了学习率和邻域函数变量。利用matlab2011软件进行求解，其中5个旅行商问题实例的结果优于MSTSP和SETSP算法，另外，10个实例的平均误差为1．4456％。实验结果表明，新算法的误差更小，并保持了SOM算法较低的计算复杂度。%There are no ellective corresponding solutions to the traveling salesman problems （TSP）. To address the problems, a new algorithm by improving learning rate and neighborhood function variance of the self-organizing map （SOM） is presented. The solutions of five instances are better than MSTSP and SETSP in matlab2011. The average error is only 1.445 6 % of anthor ten. The new algorithm has the advantage of smaller error and maintaining low SOM algorithm computational complexity.