通过建立基于交叉口关联度的协调控制子区划分模型,设计最佳控制子区划分方案获取流程,实现了交通信号控制子区的动态划分;针对最佳控制子区划分方案求解过程中可能面临的维数灾难问题,提出利用降维处理与遗传算法两种方法进行子区划分方案的快速寻优,并给出了一套协调控制子区快速动态划分流程.通过算例分析说明,在一定规模路网之下本文提出的控制子区快速动态划分方法将以一个较大概率搜索到次优子区划分方案,且因其计算时间可以被有效控制,因此完全能够满足控制子区划分的实时性要求,可推广应用于城市区域交通信号实时控制系统中.
In this paper, dynamic division of coordinated control subarea is realized by establishing the division model of control subarea with correlation degree theory and designing the process of searching for optimal subarea division scheme. Aiming at the problem of dimension disaster probablely existing in searching for optimal subarea division scheme, we present dimension-reduced processing and genetic algorithm to realize the fast optimization of subarea division. A fast dynamic division process of coordinated control subarea is proposed in detail. The example demonstrates that the fast dynamic division method of control subarea can search a sub-optimal solution with high probability under a certain scale network, and finish the optimization process in designated time. Therefore, this fast dynamic division method can satisfy the real-time request of control subarea division effectively, and can be applied to the real-time urban traffic signal control system.