An eco-driving algorithm for trains through distributing energy: A Q-Learning approach. (March 2022)

• Record Type: Journal Article
• Authors: Zhu, Qingyang; Su, Shuai; Tang, Tao; Liu, Wentao; Zhang, Zixuan; Tian, Qinghao
• Is Part Of: ISA transactions. Volume 122(2022)
• Page Start: 24
• Page End: 37
• Publication Date: 2022-03
• Subjects: 00-01 -- 99-00; Eco-driving -- Q-Learning -- Driving strategy; Engineering instruments -- Periodicals; Engineering instruments; Periodicals; Electronic journals; 629.805
• DOI: 10.1016/j.isatra.2021.04.036
• Languages: English
• ISSNs: 0019-0578

Abstract

Abstract: The energy-efficient train operation methodology is the focus of this paper, and a Q-Learning-based eco-driving approach is proposed. Firstly, the core idea of energy-distribution-based method (EDBM) that converts the eco-driving problem to the finite Markov decision process is presented. Secondly, Q-Learning approach is proposed to determine the optimal energy distribution policy. Specifically, two different state definitions, i.e., trip-time-relevant (TT) and energy-distribution-relevant (ED) state definitions, are introduced. Finally, the effectiveness of the proposed approach is verified in a deterministic and a stochastic environment. It is also illustrated that TT-state approach takes about 20 times more computation time compared with ED-state approach while the space complexity of TT-state approach is nearly constant. The hyperparameter sensitivity analysis demonstrates the robustness of the proposed approach. Highlights: The inverse problem of energy-efficient train control problem is formulated. A data-driven method based on Q-Learning approach is proposed. Two state definitions based on trip time and energy distribution are introduced.