| 强化学习型匝道控制模型的研究 |
| |
| 引用本文： | 王兴举,宫城俊彦.强化学习型匝道控制模型的研究[J].石家庄铁道学院学报,2010,23(2):104-108. |
| |
| 作者姓名： | 王兴举 宫城俊彦 |
| |
| 作者单位： | 王兴举(石家庄铁道大学,交通运输学院,河北,石家庄,050043;河北省交通安全与控制重点实验室,河北,石家庄,050043);宫城俊彦(日本东北大学,信息科学研究科,日本,仙台,9808578) |
| |
| 基金项目： | 国家自然科学基金项目，教育部留学回国人员科研启动基金 |
| |
| 摘 要： | 高速公路上的交通堵塞造成了道路利用效率低下，并伴随着能源消耗和环境污染问题，因此各种各样的高速公路控制方法应用于缓解交通堵塞。本文提出强化学习型匝道控制模型，该模型以交通流模拟为预测工具，以人工智能的强化学习为最优化选择模型，并具有一定的自主性、有记忆功能和性能反馈功能，且是一种动态的过程。应用JAVA针对不同的交通状态进行模拟再现，模拟结果表明匝道控制模型对于减少交通堵塞具有显著的效果。
|
| 关 键 词： | 匝道控制 强化学习 跟驰理论 车道变更 交通模拟 |
Reinforcement Learning Ramp Metering |
| |
| Institution: | Wang Xingju, Miyagi Toshihiko ( 1. School of Traffic and Transportation, Shijiazhuang Tiedao University, Shijiazhuang 050043, China; 2. Traffic Safety and Control Laboratory of Hebei Province, Shijiazhuang 050043, China; 3. Graduate School of Information Sciences, Tohoku University, Sendai 9808578, Japan) |
| |
| Abstract: | Since the traffic congestion in a highway as an increase in energy consumption and environmental brings about an efficiency fall of road operation as well pollution, various kinds of traffic control have been considered for easing traffic congestion until now. In this paper, reinforcement learning model is introduced. By combining this model with a simulation model for describing the traffic flow behavior in the merging sections in highways, a novel reinforcement learning ramp metering model is proposed. By numerical simulation experiments, this model shows that the effect of the proposed control measure is effective in the highway. |
| |
| Keywords: | ramp metering reinforcement learning car following model lane change model traffic flow simulation |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
