基于进化多目标优化和蚁群算法的交通控制与诱导系统研究
发布时间：2018-08-21 14:45
【摘要】：随着交通流的日益增大及复杂化,城市路网拥堵问题越来越严重,现有的智能交通控制难以提高城域交通系统的整体效率。将主动引导交通流、均衡交通资源的诱导系统和被动疏导交通流的控制系统有机结合是解决城市交通问题的有效途径。如何构建这样的智能交通系统优化模型及其优化算法成为当前的研究热点和关键技术。针对现有交通控制系统存在的问题,本文采用了将基于进化多目标优化的控制与基于蚁群算法优化的诱导有机结合的交通调控模型及其优化方法,构建了单路口多目标优化控制模型、路口间的协调机制及车辆诱导模型,能有效均衡交通负载,提高城域路网的交通效率。主要研究工作如下:(1)针对现有交通控制系统难以有效兼顾各种指标及根据实时交通状态高效调节控制信号配时方案,构建了单路口多目标优化控制模型,采用改进的进化多目标优化算法实现交通信号优化。路口控制优化模型以单位时间内通过的车辆数尽可能的多、一个周期内的平均时耗尽可能的少为优化目标。该模型可以根据实时的道路车流量信息,高效地调节自身配时方案,并能给交通决策者提供多种偏好的配时方案。为了适应交通控制系统多目标优化的需求,提出了一种多子种群并行进化的非支配排序多目标优化算法,仿真测试实验表明,该算法具有较高的时效性,较强的对pareto前沿面的探索能力和保持种群多样性的能力。(2)针对现有区域多路口协调方式中,各路口控制耦合度高,协调控制复杂,实时性差,并且对路口拥堵预判能力差等问题,构建了多路口协调控制机制,该机制通过调节路口间的车流量与道路饱合车流量的比值,来协调多个路口的运行。根据该协调机制的特点,采用了模糊控制技术进行实现。仿真验证实验表明,该协调机制能减少交通拥堵的响应时间,快速协调各个路口的信号控制,提高区域交通效率。(3)针对现有诱导系统较少考虑道路上的动态代价和出行者的起始地与目的地等问题,构建了基于多种指标的车辆诱导模型,并采用改进的蚁群算法实现对出行路径的规划。车辆诱导模型优化指标由三部分组成:起始地与目的地间的静态路径长度、该路径上通过路口总的延时转换得到的等效代价、在道路上运行时产生的动态代价。该优化模型在力求用户路径最优的同时,能尽量实现道路车辆的均衡分布。为了满足诱导系统路径规划的需求,提出了一种有偏好的蚁群算法,该算法通过偏好的设置和局部最优跳出机制,提高了全局收索能力和效率,仿真测试实验验证了算法对诱导系统路径寻优有较高效能。
[Abstract]:With the increasing and complication of traffic flow, the problem of urban road network congestion is becoming more and more serious, and the existing intelligent traffic control is difficult to improve the overall efficiency of urban transportation system. It is an effective way to solve urban traffic problems by combining active guiding traffic flow, balancing the guidance system of traffic resources with the control system of passive traffic flow. How to build such an intelligent transportation system optimization model and its optimization algorithm has become the current research hotspot and key technology. Aiming at the problems existing in the existing traffic control systems, this paper adopts a traffic control model and its optimization method, which combines evolutionary multi-objective control with ant colony optimization. The multi-objective optimal control model of single intersection the coordination mechanism between intersections and the vehicle guidance model can effectively balance the traffic load and improve the traffic efficiency of the metropolitan road network. The main research work is as follows: (1) in view of the existing traffic control system is difficult to take into account all kinds of indicators effectively and according to the real-time traffic state efficient regulation control signal timing scheme, a multi-objective optimal control model for a single intersection is constructed. An improved evolutionary multi-objective optimization algorithm is used to optimize traffic signals. The optimization model of intersection control is based on the maximum number of vehicles passing through the unit time and the possible decrease of the average time depletion in a period. Based on the real-time traffic flow information, the model can efficiently adjust its own timing scheme and provide traffic decision makers with a variety of preferred timing schemes. In order to meet the needs of multi-objective optimization in traffic control systems, a multi-objective optimization algorithm with parallel evolution of multi-subpopulations is proposed. The simulation results show that the algorithm has high time-efficiency. Strong ability to explore the pareto frontier and maintain population diversity. (2) in the existing regional multi-intersection coordination mode, the intersection control coupling degree is high, coordination control is complex and real-time is poor. To solve the problem of poor predetermination ability of traffic congestion, the coordinated control mechanism of multi-intersection is constructed. The mechanism coordinates the operation of multiple intersections by adjusting the ratio of traffic flow between intersections and traffic flow. According to the characteristics of the coordination mechanism, fuzzy control technology is adopted. The simulation results show that the coordination mechanism can reduce the response time of traffic congestion and quickly coordinate the signal control of each intersection. (3) considering the dynamic cost on the road and the origin and destination of the travelers, a vehicle guidance model based on multiple indexes is constructed. The improved ant colony algorithm is used to plan the travel path. The optimization index of the vehicle guidance model consists of three parts: the static path length between the initial location and the destination, the equivalent cost obtained by the total intersections delay conversion on the route, and the dynamic cost when running on the road. The optimization model can achieve the equilibrium distribution of road vehicles as much as possible while striving for the optimal path of the user. In order to meet the needs of path planning of induced systems, a preferred ant colony algorithm is proposed, which improves the ability and efficiency of global cable collection through preference setting and local optimal jump out mechanism. The simulation results show that the algorithm is effective for path optimization of induced systems.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U495;TP18
本文编号：2196073
[Abstract]:With the increasing and complication of traffic flow, the problem of urban road network congestion is becoming more and more serious, and the existing intelligent traffic control is difficult to improve the overall efficiency of urban transportation system. It is an effective way to solve urban traffic problems by combining active guiding traffic flow, balancing the guidance system of traffic resources with the control system of passive traffic flow. How to build such an intelligent transportation system optimization model and its optimization algorithm has become the current research hotspot and key technology. Aiming at the problems existing in the existing traffic control systems, this paper adopts a traffic control model and its optimization method, which combines evolutionary multi-objective control with ant colony optimization. The multi-objective optimal control model of single intersection the coordination mechanism between intersections and the vehicle guidance model can effectively balance the traffic load and improve the traffic efficiency of the metropolitan road network. The main research work is as follows: (1) in view of the existing traffic control system is difficult to take into account all kinds of indicators effectively and according to the real-time traffic state efficient regulation control signal timing scheme, a multi-objective optimal control model for a single intersection is constructed. An improved evolutionary multi-objective optimization algorithm is used to optimize traffic signals. The optimization model of intersection control is based on the maximum number of vehicles passing through the unit time and the possible decrease of the average time depletion in a period. Based on the real-time traffic flow information, the model can efficiently adjust its own timing scheme and provide traffic decision makers with a variety of preferred timing schemes. In order to meet the needs of multi-objective optimization in traffic control systems, a multi-objective optimization algorithm with parallel evolution of multi-subpopulations is proposed. The simulation results show that the algorithm has high time-efficiency. Strong ability to explore the pareto frontier and maintain population diversity. (2) in the existing regional multi-intersection coordination mode, the intersection control coupling degree is high, coordination control is complex and real-time is poor. To solve the problem of poor predetermination ability of traffic congestion, the coordinated control mechanism of multi-intersection is constructed. The mechanism coordinates the operation of multiple intersections by adjusting the ratio of traffic flow between intersections and traffic flow. According to the characteristics of the coordination mechanism, fuzzy control technology is adopted. The simulation results show that the coordination mechanism can reduce the response time of traffic congestion and quickly coordinate the signal control of each intersection. (3) considering the dynamic cost on the road and the origin and destination of the travelers, a vehicle guidance model based on multiple indexes is constructed. The improved ant colony algorithm is used to plan the travel path. The optimization index of the vehicle guidance model consists of three parts: the static path length between the initial location and the destination, the equivalent cost obtained by the total intersections delay conversion on the route, and the dynamic cost when running on the road. The optimization model can achieve the equilibrium distribution of road vehicles as much as possible while striving for the optimal path of the user. In order to meet the needs of path planning of induced systems, a preferred ant colony algorithm is proposed, which improves the ability and efficiency of global cable collection through preference setting and local optimal jump out mechanism. The simulation results show that the algorithm is effective for path optimization of induced systems.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U495;TP18
【参考文献】
相关期刊论文 前3条
1 陈昕;杨兆升;王海洋;刘新杰;王彦新;;城市交通控制与诱导系统协同研究[J];公路交通科技;2007年04期
2 秦敏;;道路交通瓶颈路口的红波协调控制策略[J];贵州师范学院学报;2010年06期
3 高云峰;胡华;陈红洁;杨晓光;;交叉口群交通控制实时评价模型仿真研究[J];系统仿真学报;2007年24期
，本文编号：2196073
本文链接：https://www.wllwen.com/kejilunwen/daoluqiaoliang/2196073.html
