用Hopfield神经网络与遗传算法求解TSP问题的实验比较与分析

Hopfield neural network and genetic algorithm in solving travelling salesman problem: experimental comparison and analysis

首先介绍了实验中利用 Hopfield神经网络和遗传算法求解 TSP问题的程序设计方法 ,分析了 Create Chromosome算法的时间复杂性 ,并测试了群体规模对解最优性的影响 .从计算方法的软件实现角度出发 ,比较了两种计算技术的相似点与不同之处 .本次实验的结论是 :遗传算法比Hopfield神经网络求解 TSP问题的效率高 ,且随着问题规模的扩大 ,优势更为突出 .

This paper firstly introduces programmes for HopfieldTSP and GeneticTSP and describes consequently the chromosome algorithm in detail, which concentrates on time complexity and effects of population size on optimized path. Comparison results and operation efficiency between the HopfieldTSP and the GeneticTSP algorithms, this paper supposes that the genetic algorithm is superior to the Hopfield neural network algorithm in solving the travelling salesman problem.