中文摘要：

面向复杂多变的交通系统控制需求，提出一种考虑交通管理策略的交叉口信号控制多目标优化模型及算法，其步骤为：首先，构建城市道路交叉口多目标优化模型；然后，引入交通管理者的交通控制策略，基于模糊分析法确定各优化指标权重；最后，考虑Q学习算法简单方便且具有快速收敛性，基于Q学习算法对多目标优化模型进行求解，从而实时产生考虑交通管理者策略的交叉口信号控制方案。仿真结果表明，所提出的方法不仅能够充分反映交通管理者的控制策略，而且能够有效提高交叉口通行效率，相比传统方法具有缩短排队长度、降低延误时间和减少停车次数的优势，并且这种优势随着交通流量的增大而更加明显。

英文摘要：

A multi-objective optimization model and its algorithm aimed at the intricate demand of traffic system controlling were presented. The procedures were as follows. The traffic controlling strategy was drawn into the model after the model was built. The weight of optimization index was determined based on fuzzy analytic process. The solutions of the multi-objective optimization model were obtained by Q-algorithm, based on the fact that the Q-algorithm is fast constrained and convenient. The program of intersection signal controlling based on traffic management strategy was real-time produced. The simulation results show that the method can not only fully reflect the control strategies of traffic managers, but also improve the travelling efficiency at intersection. Compared with the traditional method, it has some advantages such as the queue length shortened, the time delayed and the number of parking reduced. These advantages are more obvious with the increase of the traffic flow.