Traffic Signal Control with Cell Transmission Model Using Reinforcement Learning for Total Delay Minimisation

摘要

This paper proposes a new framework to control the traffic signal lights by applying the automated goal-directed learning and decision making scheme, namely the reinforcement learning (RL) method, to seek the best possible traffic signal ac- tions upon changes of network state modelled by the signalised cell transmission model (CTM). This paper employs the Q-learning which is one of the RL tools in order to find the traffic signal solution because of its adaptability in finding the real time solu- tion upon the change of states. The goal is for RL to minimise the total network delay. Surprisingly, by using the total network delay as a reward function, the results were not necessarily as good as initially expected. Rather, both simulation and mathemat- ical derivation results confirm that using the newly proposed red light delay as the RL reward function gives better performance than using the total network delay as the reward function. The investigated scenarios include the situations where the summa- tion of overall traffic demands exceeds the maximum flow capacity. Reported results show that our proposed framework using RL and CTM in the macroscopic level can computationally efficiently find the proper control solution close to the brute-forcely searched best periodic signal solution (BPSS). For the practical case study conducted by AIMSUN microscopic traffic simulator, the proposed CTM-based RL reveals that the reduction of the average delay can be significantly decreased by 40% with bus lane and 38% without bus lane in comparison with the case of currently used traffic signal strategy. Therefore, the CTM-based RL algorithm could be a useful tool to adjust the proper traffic signal light in practice.