摘要
为解决大规模区域交通控制滚动优化问题中的约束条件复杂、解空间规模庞大的最优化难题,提出了一种基于改进蚁群算法的降阶滚动优化算法.基于宏观交通流模型建立了区域交通控制滚动优化模型,在蚁群算法中设计了层状解构造图对该模型解空间进行描述和求解.运用降阶方法将大规模区域分解成一系列子区域,在蚁群算法中设计了复合层状解构造图对该降阶模型的解空间进行描述和求解,并分析了基于两种解构造图的蚁群算法的计算复杂度.分析和仿真结果表明,该降阶算法提高了整体计算效率,明显地降低了总停车延误时间,适用于大规模区域交通控制的滚动优化.
To deal with the optimizing difficulties of complex constraints and large-scale solution space in the rolling horizon optimization problem of large-scale regional coordinated traffic control, a reduced-order rolling horizon optimization algorithm based on an improved ant algorithm was proposed. A rolling horizon optimization model of regional traffic control problem was formulated based on a macroscopic traffic model, then a layered construction graph in the ant algorithm was designed to describe and search the solution space of the model. A large-scale region was divided into a series of subregions with a reduced-order method, and a compound layered construction graph was designed to describe and search the solution space of the reduced-order model. The computation complexity of the two algorithms based on different construction graphs was analyzed. Analysis and simulation results show that the reduced-order algorithm improves the total computational efficiency, reduces the total stop delay remarkably, and is suitable for the rolling horizon optimization of large-scale traffic regions.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2005年第6期835-839,848,共6页
Journal of Zhejiang University：Engineering Science
基金
国家"973"重点基础研究发展规划资助项目(2002CB31220303).
关键词
交通控制
蚁群算法
滚动优化
Algorithms
Computational complexity
Optimization
