基于Sarsa算法的城轨列车节能控制策略研究OA北大核心

Focusing on the problem of energy-efficient operation of urban rail transit,an energy-efficient control strategy of urban rail transit based on Sarsa reinforcement learning algorithm is proposed.To achieve the driving strategy of reducing energy consumption while taking into account punctuality and comfort when the urban rail transit is in automatic driving mode and facing different road conditions,the running state of the train is discretized and the continuous driving process is divided into several sub-sections according to the line conditions.Combined with the speed limit,initial state and terminal state of the track,the appropriate reward functions are constructed based on energy consumption and running time.At the same time,the setting of optional speed is limited by the maximum speed and minimum speed that can be reached in the current state,which reduces the exploration space and accelerates the convergence of the algorithm.Finally,the simulation is carried out for the case of Xiaohongmen-Xiaocun Station on the Yizhuang Urban Rail Line in Beijing.The experimental results show that,compared with the traditional dynamic programming method,the Sarsa algorithm saves 9.32%energy while meeting the requirements of comfort and punctuality.Compared with the Q-learning algorithm in reinforcement learning,the number of overspeed also decreases significantly in the process of speed selection.Simulation results show that the Sarsa algorithm has a better energy-saving effect and security.With the algorithm parameters unchanged,the speed limit conditions are adjusted,and compared with the traditional dynamic programming method again,it still saves 4.21%energy,which verifies the robustness of the algorithm.