Abstract: A multi-objective traffic control model for single intersection is proposed，in which vehicle delay，stopping time and traffic capacity are taken as performance indexes，and saturation degree is taken as restrictions.The method can adjust the weight of performance indexes according to traffic state.The objective model is solved by genetic algorithm.The result shows that the model can adapt the traffic state well and improve the traffic congestion effectively.

摘要： 建立了在饱和度约束下以延误、停车次数、通行能力为性能指标的单交叉口多目标控制模型；模型可以根据交通状态的不同对交通控制的目标有所侧重：交通状态趋向拥挤时要确保发挥路口的最大通行能力，非拥挤状态时应尽可能减少路口信号控制的延误和停车次数；最后利用遗传算法对模型进行优化。实验结果表明，该模型能较好地适应实际的交通状态的变化，可有效改善交叉口的交通状况。