SOFM神经网络最近插入法混合算法在TSP问题中应用研究

The Applied Research of SOFM Neural Networks-The Nearest Insertion Hybrid Algorithm in the TSP Problem

SOFM神经网络已经成功应用到TSP问题中,但是该算法存在一些缺点,随着学习速度逐步降低,会导致一些城市无法通过。针对这些缺点,尝试在SOFM神经网络中引入最近插入法形成混合算法。通过实验,并与SOFM神经网络该算法对比,结果表明,该算法能够很好地完善该问题。

SOFM neural network has been successfully applied to the TSP problem, but the algorithm has some shortcomings, as the learning rate gradually reduee, will result in some eities not pass through. In response to these shortcomings, the Nearest Insertion in the SOFM neural network was introduced to form a hybrid algorithm. Through the experiment, and with the SOFM neural network the algorithm comparison, the results show that the algorithm is well positioned to improve the problem.