城市道路交叉口群信号协调控制多目标优化与仿真

Multi-objective Optimization and Simulation for Urban Road Intersection Group Traffic Signal Control

为克服采用单目标或线性组合方法研究城市路网交通信号协调控制与优化的弊端,以交叉口群为对象,运用多目标优化方法建立了交通信号协调控制参数的多目标优化模型,并基于非支配排序遗传算法设计了相应的多目标优化算法求解模型。采用改进的元胞传输模型对信号控制路网上交通流的运行与演化过程进行仿真模拟并获取各个优化目标值。分析结果表明:单目标优化模型使被优化目标最优的同时,往往使其他未被优化的1个或几个目标较差;线性组合优化模型可获得折衷解;多目标优化方法能更好地兼顾多个优化目标,使最优或较优的目标数多于较差的目标数。

To overcome the limitations of the linear combination and single objective optimization method,a multi-objective optimization and simulation for intersection group traffic signal control were proposed.A multi-objective optimization algorithm based on non-dominated sorting genetic algorithm Ⅱ was provided to solve the model.The dynamic traffic flow on signalized urban road network was simulated and evaluated by an enhanced cell transmission model（CTM） to find out the values of optimized objectives.The simulation results show that the single objective optimization method makes some other objectives worse when the objective optimized reachs optimal value,and the linear combination optimization method can gain a compromised solution,but the multi-objective optimization method gives consideration to more objectives,making the number of optimal or suboptimal ones more than that of worse ones.