基于约束的Hopfield网络优化TSP问题研究

The research of constraint-based Hopfield network application to TSP

TSP问题是一类典型的组合优化NP问题,在传统Hopfield神经网络的基础上增加了四个条件约束,通过求取条件约束的最小值而得到目标问题的最优或次优解,并推导证明了算法的收敛性,同时通过选取适当的运行参数及阈值函数在实例中验证了算法的有效性.

TSP is a typical combinatorial optimization NP Hard problems, providing a new algorithm by adding four conditions bound based on traditional Hopfield neural networks, which can get the target problem＇s optimal or suboptimal solution bound through getting the minimum value of the conditions bound. The convergence of the algorithm was also proved. So when applied to TSP problem, this algorithm is more efficient to obtain the optimal solution or suboptimal solution.