arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1911.04400cs.AIcs.LG
收藏

用模型驱动强化学习

Driving Reinforcement Learning with Models

imperial.ac.uk(帝国理工学院(注:原文本中此处可能为邮箱域名,非明确机构,根据提取规则应忽略))

Meghana Rathi, Pietro Ferraro, Giovanni Russo

摘要:

在本文中,我们提出了一种新的方法来补充强化学习(RL)与基于模型的控制(特别是,模型预测控制- MPC)。我们介绍了一种算法,MPC增强RL(MPRL),结合RL和MPC在一个新的方式,使他们可以增强对方的优势。我们证明了MPRL的有效性,让它发挥对雅达利游戏乒乓。对于这项任务,结果强调了MPRL如何能够在单独使用时优于RL和MPC。

英文摘要:

In this paper we propose a new approach to complement reinforcement learning (RL) with model-based control (in particular, Model Predictive Control - MPC). We introduce an algorithm, the MPC augmented RL (MPRL) that combines RL and MPC in a novel way so that they can augment each other's strengths. We demonstrate the effectiveness of the MPRL by letting it play against the Atari game Pong. For this task, the results highlight how MPRL is able to outperform both RL and MPC when these are used individually.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑