The embodiment of the invention provides a train energy-saving operation control method based on safety reinforcement learning and a related device. The method comprises the following steps: establishing a distance discretization train energy-saving operation simulation environment, resetting unsafe actions through a safe action resetting mechanism in network training in order to meet the constraint of speed limitation in the train operation process, and establishing a distance discretization train energy-saving operation simulation model by taking the minimum energy consumption as a target under the constraint of planned operation time. A relaxed dynamic reward function is designed to guide continuous evolution of the network to convergence. On the real train data and route information, it is verified that the operation control strategy output by the method meets the expectation, the requirement for the planned operation time of the train can be met, and meanwhile the purpose of the optimal energy consumption is achieved. On the basis, the training speed of the optimal strategy can be increased, and the control precision of the control strategy on train operation is guaranteed.
本发明实施例提供了一种基于安全强化学习的列车节能运行控制方法及相关装置。其中,方法包括:建立距离离散化的列车节能运行仿真环境,为满足列车运行过程中速度限制的约束,在网络训练中,通过一种安全动作重置机制来对不安全动作进行重置,考虑计划运行时间的约束下,以能耗最小为目标,设计了一种松弛的动态奖励函数,引导网络的持续进化至收敛。在真实的列车数据和线路信息上,验证了方法输出的运行控制策略符合预期,能够满足列车计划运行时间的要求,同时达到能耗最优的目标。基于此,本发明实施例能够加快最优策略的训练速度,保证了控制策略对列车运行的控制精度。
Train energy-saving operation control method based on safety reinforcement learning and related device
基于安全强化学习的列车节能运行控制方法及相关装置
2024-06-28
Patent
Electronic Resource
Chinese
IPC: | B61L Leiten des Eisenbahnverkehrs , GUIDING RAILWAY TRAFFIC |
Train energy-saving auxiliary driving method and system based on safety reinforcement learning
European Patent Office | 2024
|European Patent Office | 2025
|European Patent Office | 2023
|European Patent Office | 2023
|