一种具有连续滞后神经元的Hopfield网络求解TSP问题的算法(英文)

• 作 者: 张威
• 机 构: 浙江海洋学院数理与信息学院,
• 出 处: 浙江海洋学院学报(自然科学版)
• 发表日期: 2013年01期
• 关键词: 神经网络 TSP问题 旅行商问题 Hopfield网络 hysteresis 连续滞后 仿真结果 计算效率 计算性能 iteration

提出了一种具有连续滞后神经元的Hopfield神经网络体系结构,有效地解决旅行商问题(TSP)。对连续滞后神经元的Hopfield神经网络结构和整个网络的计算性能进行了研究。理论上通过模拟随机生成的具有连续滞后神经元的Hopfield神经网络证明网络的优越性。通过模拟TSP问题的应用来说明此算法的计算效率。仿真结果表明,具有连续滞后神经元的Hopfield神经网络在解决TSP问题时,无论从处理时间和解决质量上来看都优于传统的Hopfield神经网络和二元连续Hopfield神经网络。

A Hopfield neural network architecture with continuous hysteresis neurons is proposed to effectively solve the traveling salesman problem (TSP). The Hopfield neural network structure of continuous lag neurons and the computational performance of the whole network are studied. In theory, Hopfield neural networks with continuous hysteresis neurons generated randomly are simulated to prove the superiority of the network. The computational efficiency of this algorithm is illustrated by simulating the application of the TSP problem. Simulation results show that Hopfield neural networks with continuous hysteresis neurons are superior to traditional Hopfield neural networks and binary continuous Hopfield neural networks in terms of processing time and quality of solution in solving TSP problems.