Multi-intersection signal adaptive control method based on reinforcement learning

Abstract

The invention discloses a multi-intersection signal adaptive control method based on reinforcement learning, and the method comprises the steps: dividing continuous time into discrete time intervals according to the actual conditions of a plurality of intersections to be controlled, and building a random traffic model in combination with a state transition model and a traffic mechanism model; defining state, action, reward and value functions of reinforcement learning, and establishing a multi-intersection control model according to a random traffic model; a simulation intersection environment and traffic flow data are configured to a traffic simulator, and an intelligent agent network based on a reinforcement learning framework is built; and respectively transmitting the real-time traffic information into the intelligent agent networks corresponding to the intersections according to the positions of the intersections to obtain phases of the intersections output by the intelligent agent networks, and executing traffic signal control of the intersections according to the phases of the intersections. The method improves the precision of the model and the control efficiency of the intersection signal, facilitates the alleviation of traffic congestion, and can be widely applied to the technical field of traffic signal control.

Description

technical field [0001] The invention relates to the technical field of traffic signal control, in particular to a multi-intersection signal adaptive control method based on reinforcement learning. Background technique [0002] Urban traffic has a leading and overall impact on a city's economic, cultural and environmental development. With the advancement of economic prosperity and urbanization, the number of automobiles and transportation demand continue to surge, and the construction of urban infrastructure is difficult to keep up with the growth of vehicle development and transportation demand. In addition, problems such as unreasonable urban planning and land use, insufficient public transportation capacity or unreasonable route layout lead to frequent traffic congestion. Improving the traffic capacity of the road network, alleviating traffic congestion, and ensuring the orderly development of urban traffic are urgent issues facing urban traffic control. Urban traffic c...

AI Technical Summary

Problems solved by technology

Most of these methods assume that the traffic conditions are known, thus ignoring the randomness of the traffic system

Moreover, when the reinforcement learning algorithm based on the value function is applied to the control of multi-intersection signal lights, its complexity increases exponentially with the increase of the state space and action space, and it faces the curse of dimensionality.

Images