Traffic optimization of road network agents based on multi-attribute fuzzy decision-making

Abstract

The application of multi-agent reinforcement learning in the coordinated control of urban road network traffic has greatly alleviated the local congestion in urban traffic. However, most of the current researches use a single performance index of the road network for optimal control, which will inevitably lead to uncoordinated traffic on the road network and congestion in local sections. Aiming at the above problems, based on the multi-agent reinforcement learning algorithm, this paper comprehensively considers the average delay and queuing length of the road section, and proposes the coordinated control of multi-attribute interval decision-making, so that the traffic status of each road section in the road network can be optimized. The control strategy is simulated and verified by the simulation software SUMO . The simulation results show that compared with the traditional timing control, the control method proposed in this paper can greatly reduce the average vehicle delay and queue length, reduce the pressure of local congestion, and improve the pass efficiency.