基于混沌神经网络的最优路径选择算法

研究车载交通流诱导系统的最优路径选择问题。采用广义路阻的定义,考虑了驾驶员在路径选择中的不同要求,并借助一种具有暂态混沌和时变增益的神经网络(NNTCTG),针对最优路径选择问题设计了神经网络结构,构造了能量函数,提出了一种能够满足不同出行者偏好的最优路径选择算法。所提出的算法具有很多优良特性,即暂态混沌特性和平稳收敛性,能有效地避免传统Hopfield神经网络极易陷入局部极值的缺陷。它通过短暂的倒分叉过程,能很快进入稳定收敛状态。仿真结果表明,NNTCTG求解指定起讫点对之间的最优路径问题时,总能收敛到全局最优,同时具有更高的搜索效率。

Optimal Route Selection Algorithm Based on Chaotic Neural Networks

The issue about optimal route selection of an in-vehicle traffic flow guidance system was discussed.By using the definition of general road weight,different kinds of requirements were fully considered.Based on a neural network with transient chaos and time-variance gain,neural network structure and computational energy function for solving optimal route selection was constructed,and then an algorithm which can support the driver in deciding on an optimal route to his preference was proposed.The proposed algorithm has many merits such as transient chaos and stable convergence so as to overcome the drawbacks of easily getting stick in local extremum in conventional Hopfield neural networks.It can reach a stable convergent state after shortly reversed bifurcation.Simulation results show that using NNTCTG in searching the optimal route between an appointed origin-destination pair always converges to the globally optimal solution and it has higher efficiency of searching than HNN.