Hopfield网络解旅行商问题的动态消元算法

对Hopfield网络解旅行商问题的算法做了改进工作:为了消除无效解,给出了普适的初始状态,调整了差分迭代方程的参数,改变了稳定状态的判断.为了得到最优解,提出动态消元算法,要求消元后第r次的运算结果优于或等于第r-1次的运算结果.计算机模拟表明此算法对初始条件具有稳健性,从任何初始状态开始都能得到最优解.

Dynamic elimination method for solving TSP by Hopfield network

An improved algorithm to solve TSP by the Hopfield network was presented. For avoiding the invalid solution, a universal initial state was given, the parameters of difference iterative equations were adjusted and the criterion of stability was changed. For yielding the optimal solution, a dynamic elimination method which claims the rth result must shorter than or equal to the (r-1)th result is established. Computer simulations showed that this method is not sensitive to initial conditions at all and allows us to obtain the optimal solution starting from arbitrary initial state.