arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2403.09110cs.LGcs.SYeess.SYmath.DSmath.OC
收藏

SINDy-RL:可解释且高效的基于模型的强化学习

SINDy-RL: Interpretable and Efficient Model-Based Reinforcement Learning

Department of Mechanical Engineering, University of Washington, Seattle, WA 98195, USA(华盛顿大学机械工程系);Data Science and Artificial Intelligence Department, The Aerospace Corporation, El Segundo, CA 90245(航空航天公司数据科学与人工智能部);Department of Aeronautics, Imperial College, London SW7 2AZ, United Kingdom(帝国理工学院航空系);Department of Applied Mathematics, University of Washington, Seattle, WA 98195(华盛顿大学应用数学系);Department of Electrical and Computer Engineering, University of Washington, Seattle, WA 98195(华盛顿大学电气与计算机工程系)

Nicholas Zolman, Christian Lagemann, Urban Fasel, J. Nathan Kutz, Steven L. Brunton

AI总结:

本文提出SINDy-RL框架,结合SINDy和DRL,实现低数据下高效、可解释的动力学模型和控制策略,通过基准环境和流体控制实验验证其有效性。

摘要:

深度强化学习(DRL)在复杂环境中揭示复杂控制策略方面展现出巨大潜力,如稳定托卡马克聚变反应堆或最小化流体中物体的阻力。然而,DRL需要大量训练示例且成本高昂。此外,依赖深度神经网络导致不可解释的黑箱策略,可能在嵌入式系统中计算成本过高。最近的稀疏字典学习进展,如非线性动态的稀疏识别(SINDy),在低数据条件下展示了创建高效且可解释的数据驱动模型的潜力。本文介绍SINDy-RL,一种结合SINDy和DRL的统一框架,以创建高效、可解释且可信的动力学模型、奖励函数和控制策略。我们在基准控制环境和流体控制问题上展示了方法的有效性,包括在Re=1000时的3D NACA 0012翼型气流抑制。SINDy-RL在显著较少的环境交互中实现了与现代DRL算法相当的性能,并产生比DRL策略小多个数量级的可解释控制策略。

英文摘要:

Deep reinforcement learning (DRL) has shown significant promise for uncovering sophisticated control policies that interact in complex environments, such as stabilizing a tokamak fusion reactor or minimizing the drag force on an object in a fluid flow. However, DRL requires an abundance of training examples and may become prohibitively expensive for many applications. In addition, the reliance on deep neural networks often results in an uninterpretable, black-box policy that may be too computationally expensive to use with certain embedded systems. Recent advances in sparse dictionary learning, such as the sparse identification of nonlinear dynamics (SINDy), have shown promise for creating efficient and interpretable data-driven models in the low-data regime. In this work we introduce SINDy-RL, a unifying framework for combining SINDy and DRL to create efficient, interpretable, and trustworthy representations of the dynamics model, reward function, and control policy. We demonstrate the effectiveness of our approaches on benchmark control environments and flow control problems, including gust mitigation on a 3D NACA 0012 airfoil at $Re=1000$. SINDy-RL achieves comparable performance to modern DRL algorithms using significantly fewer interactions in the environment and results in an interpretable control policy orders of magnitude smaller than a DRL policy.

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。

↑