高速公路入口匝道模糊神经网络ACO控制

Ant Colony Optimization Control for Fuzzy Neural Network in Freeway Entrance Ramp

针对入口匝道控制中局部需求大于高速公路主线容量情况下Alinea控制算法不能有效反馈的问题,结合模糊控制和神经网络的优点,通过神经网络来训练模糊控制规则,提出蚁群算法优化的模糊神经网络控制器,并对控制器应用于入口匝道控制进行了详细设计。仿真结果表明,基于蚁群优化算法的模糊神经网络控制器学习次数远小于Alinea控制算法,且收敛速度快,运算效率高,控制品质好,能够更好地稳定主线交通流密度。

On-ramp control is an effective way to reduce the congestion of freeway. The fuzzy control and the neural network methods both have their distinctive advantages in the on-ramp control. Based on the ant colony optimization （ACO） algorithm, the fuzzy control was combined with neural network to realize the freeway ramp metering optimized by Ant colony algorithm. The simulation for this algorithm was carried out and the results show that this algorithm is effective and the learning number of our algorithm is far less than other Alinea algorithms. It can efficiently maintain the mainline traffic flow density within a desired value range.