智能交通网络信号控制系统的自适应协同优化方法研究

项目名称： 智能交通网络信号控制系统的自适应协同优化方法研究

项目编号： No.61202342

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 计算机科学学科

项目作者： 张晓勇

作者单位： 中南大学

项目金额： 23万元

中文摘要： 本项目针对日益严峻的城市交通拥塞问题，提出一种智能交通网络信号控制系统的自适应协同优化方法，平衡区域交通负载，提高整个城市交通网络的通行效率和交通管理的可靠性和可扩展性。采用量化微分动态逻辑和分段仿射的方法建立大规模城市交通流的量化混合模型；引入性能判据函数寻求量化混合模型的交通网络优化问题，利用分支定界法简化问题描述，通过对偶分解将其转化为城市交通信号控制系统的凸优化问题；设计一种结合模型预测控制策略的在线优化分布式协同控制律，同时利用相对对角可控的方法分析系统特性，确保系统在不可预测情况下的协同控制收敛性；引入增强学习方法，利用行为/评价网络的反馈、更新和交互，建立协同控制器的输入调节机制，增强对系统不确定性和突发扰动的自适应能力。通过本项目的研究，将为城市智能交通网络的信号控制系统提供一种有效的自适应协同优化方法。

中文关键词： 智能交通系统；区域负载均衡；协同优化；模糊Q学习；交通诱导

英文摘要： Focusing on the increasingly sever urban traffic congestion problems, this study proposes an adaptive cooperative optimization scheme for signal control system in intelligent traffic networks. This scheme can be used to balance the regional traffic and enhance the traffic efficiency, the reliability and scalability of traffic management. The quantified differential dynamic logic and the piecewise affine methods are presented to build quantified hybrid model for large-scale urban traffic flows. And the performance criterion functions are introduced to obtain the optimization problem of the proposed model. A branch-and-bound method is utilized to simplify the problem, which can be transformed into a convex optimization problem by dual-decomposition. An online distributed cooperative control scheme is constructed combining model predictive control strategy. System performance is analyzed by adopting the relative diagonal amplitude dominant theory to guarantee the convergence of cooperative control under unpredictable situations. By utilizing the feedback, updating and interactions of the actor/critic networks based on reinforcement learning, an input regulator is developed for the cooperative controller. This regulator can enhance the adaptability to system uncertainties and disturbances. This study will develop an

英文关键词： Intelligent transport system；region traffic balance；cooperative optimization；fuzzy Q-learning；traffic guidance