Study On Urban Arterial Traffic Signal Optimization Based On Multi-agent Switched Control

With China's economic development entering a new era,people's living needs are constantly changing with rapid economic development,social changes and higher living standards,especially in terms of transportation.The number of automobile-owned in cities are still growing at a high speed,and traffic congestion and other problems have been caused by excessive car ownership in cities.In view of the complexity and strong dynamic and randomness of urban traffic signal control system and the limitations of conventional signal control methods,it is particularly important to improve the traffic capacity of urban roads by formulating scientific and reasonable urban intelligent traffic control methods.Therefore,regarding to the weaknesses and shortcomings of the existing urban traffic signal control methods,this dissertation combines the intelligent control theory to discuss and analyze the signal control at urban intersections and the urban arterial coordinate control.The following aspects are maded a study and analysis in depth:Firstly,the basic ideas of urban arterial coordinate control are expounded,and the control parameters and evaluation indicators of urban arterial traffic signals are introduced.The basic working principle of agent traffic signal control is described.The necessity of modeling,analyzing and researching signalized intersections is introduced based on the switched system model.Secondly,according to multi-agent technology and switched system theory,combined with the characteristics of urban intersection signal control,a closed-loop hybrid control structure of urban intersection signal based on multi-agent is constructed.On this basis,the intersection intelligent agent based on Q learning algorithm and the switched control agent based on multi-phase fuzzy phase sequence optimization are designed respectively by reference to people's decision-making processes of traffic command and control strategy for multi-phase urban intersections.Thirdly,the switched system theory is applied to the study of urban traffic control modeling,and the dynamic mathematical model of traffic flow at single intersection based on the switched system is established.The closed-loop hybrid control method is used to control the traffic flow of the switched system.The average queue length of the vehicle is used as the evaluation index,and the traffic flow situation of the intersection is simulated.and the effectiveness of the multi-phase closed-loop hybrid control method proposed in this dissertation is verified.Finally,based on the multi-phase closed-loop hybrid control of single intersection,a multi-agent system for urban arterial coordinate control is designed,and chaotic disturbance is introduced into Q-learning method to improve the agent's intelligence.A two-level hierarchical architecture based on multi-agent is constructed and applied to the urban arterial dynamic mathematical model using switched system theory.The global optimization problem in urban traffic control is transformed into finding the optimal phase switched time and conversion sequence for the hybrid switched system,so as to achieve the optimal control of urban arterial traffic.