同济大学学报（自然科学版）2024，Vol.52Issue(6)：928-934,981,8.DOI:10.11908/j.issn.0253-374x.22418
典型匝道控制场景下深度强化学习决策机理解析
Understanding Deep Reinforcement Learning Algorithm in Typical Ramp Metering Scenarios
摘要
Abstract
This paper presents the control mechanism of deep reinforcement learning(DRL)in a typical ramp metering scenario.The state value function is used to evaluate if the DRL model has the ability to distinguish the change of state.The saliency map is used to perceive the state key features and control pattern for the DRL model under specific traffic states.By using the input perturbation,the action match ratio and control performance under perturbed data are analyzed to explore the key areas of control.The results show that the DRL model can evaluate the traffic state accurately,distinguish the key features,and then make reasonable decisions.关键词
交通工程/深度强化学习/可解释机器学习/匝道控制Key words
traffic engineering/deep reinforcement learning(DRL)/explainable machine learning/ramp metering分类
交通工程引用本文复制引用
刘冰,唐钰,暨育雄,沈煜,杜豫川..典型匝道控制场景下深度强化学习决策机理解析[J].同济大学学报（自然科学版）,2024,52(6):928-934,981,8.基金项目
上海市科委科研计划(19DZ1209100) （19DZ1209100）
浙江省重点研发计划(2021C01011) （2021C01011）
