基于遗传算法的城市交叉口实时信号控制研究

Real-time Signal Control of Urban Intersection Based on Genetic Algorithm

研究了相对固定周期条件下交叉口信号配时随路上交通流量变化而实时调整的优化算法。以各相各进口道上总的延误时间最少为目标建立优化函数 ,以饱和度限制及最小绿灯时间要求为约束条件建立非线性规划模型 ,采用遗传算法对模型进行求解以得到最优配时方案。通过仿真实例说明了函数构造和求解过程。在仿真实例中比较了不考虑饱和度限制下的优化情况和仅以各相关键进口道的总延误最小为目标的优化情况。结果表明 ,为保证优化的正确性及合理性 ,应考虑最大饱和度的限制 ,并以各相各进口道的总延误最小为目标 ,建立目标优化函数。

This paper focuses on optimizing algorithms of signal timing of urban intersecti on based on real-time traffic volume under relatively fixed periods. Firstly o ptimizing function is made to minimize total delay of all approaches of all phas es. Then non-linear programming model with restrictions of saturation and minim um green light time is constructed. And then genetic algorithm is applied to sol ve this model and the best scheme of signal timing can be obtained. The process of constructing and solving the function is showed in a simulation experiment. I n this experiment the optimization with no regard of the restriction of saturati on and the optimization with the target of minimum total delay of all key approa ches of all phases are compared. The result shows that the restriction of satura tion should be considered and the goal function should be constructed from the t arget of minimum total delay of all approaches of all phases in order to ensure correctness and rationality of optimization.