Abstract: A self-adaptive signal timing optimization model is conducted for a single intersection. In the model, the traffic flux of the next cycle is forecast with one of the current cycle. The vehicle queue length of every phase when green ending is adopted as one of the multiple control objectives. All the objectives are integrated into a target in the way of ideal point. The genetic algorithm is adopted in the paper. The simulation computing data shows the algorithm is effective and speedy.

Key words: signal timing, genetic algorithm, multiple objectives, self-adaptive optimization, queue length

摘要： 建立了一种单交叉口自适应优化配时模型。通过对本周期交通流数据线性预测下一周期各车道的排队长度，以各相位绿灯结束时的排队长度最小作为优化目标，建立多目标优化函数。通过采用理想点法，运用遗传算法进行优化。仿真实验结果表明，该方法优化后的配时方案能够反映各相位实际交通需求，并具有很好的实时性。

关键词: 信号配时, 遗传算法, 多目标, 自适应优化, 排队长度

CLC Number: 

• U491.5