整合快速路网匝道控制和路径诱导的进化粒子群算法

针对快速路匝道控制和路径诱导优化，改进已有宏观交通流模型对上下匝道的处理方式，提出了一种实现快速路网协同整合动态优化控制的进化粒子群算法。利用区分目的地车流的均匀分布，给出起始路段区分目的地车流的实际驶入比例。通过对实际路网中上下匝道车流的观察分析，给出上匝道车速与下游主线车速的关联关系，并对下匝道实施类似主线路段的建模处理。针对基于上述改进得到的快速路网动态控制系统，利用控制变量的箱式约束，在经典粒子群算法中引入交叉变异操作，给出了一种高效的进化粒子群算法。通过算例分析比较了经典粒子群算法和进化粒子群算法，证实了新方法可以高效处理复杂的实际快速路网。

An evolutional particle swarm optimization applied to freeway network control with ramp metering and route guidance

Through modifying the treatment of on- and off-ramp in the classical macroscopic traffic flow model, this paper proposed an evolutional particle swarm optimization (PSO) algorithm to deal with the coordinated and integrated control of freeway network with ramp metering and route guidance. With the assumption of uniform distribution of demand with specified destination, this paper clarified the actual rate of traffic flow with specified destination at origin. Based on the observation of the actual traffic at ramps, this study built the connection between speeds at on-ramp and its downstream main line. Modeling the off-ramp as a main line segment can describe the real traffic well. To solve the modified freeway network optimal control efficiently, this study introduced the crossover and mutation operator in the classical PSO by making use of the box-type constraints of control variables. By comparing the efficiencies of the classical PSO and our evolutional PSO in the numerical analysis, this study shows that our new algorithm can efficiently deal with complex real freeway networks.