pith. sign in

arxiv: 2006.03947 · v2 · pith:OE7J54UCnew · submitted 2020-06-06 · 📡 eess.SY · cs.LG· cs.SY· stat.ML

Neural Lyapunov Redesign

classification 📡 eess.SY cs.LGcs.SYstat.ML
keywords lyapunovcontrollerstabilityalgorithmcontroldynamicaleffectivefunction
0
0 comments X
read the original abstract

Learning controllers merely based on a performance metric has been proven effective in many physical and non-physical tasks in both control theory and reinforcement learning. However, in practice, the controller must guarantee some notion of safety to ensure that it does not harm either the agent or the environment. Stability is a crucial notion of safety, whose violation can certainly cause unsafe behaviors. Lyapunov functions are effective tools to assess stability in nonlinear dynamical systems. In this paper, we combine an improving Lyapunov function with automatic controller synthesis in an iterative fashion to obtain control policies with large safe regions. We propose a two-player collaborative algorithm that alternates between estimating a Lyapunov function and deriving a controller that gradually enlarges the stability region of the closed-loop system. We provide theoretical results on the class of systems that can be treated with the proposed algorithm and empirically evaluate the effectiveness of our method using an exemplary dynamical system.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. PolarNet: Single-Minima Neural Network for Modeling Lyapunov Functions

    eess.SY 2026-05 unverdicted novelty 6.0

    PolarNet is a neural architecture that guarantees a single critical point and is proven proper and universal for modeling Lyapunov functions, serving as a drop-in fix for training issues in neural Lyapunov control.

  2. Learning to Adapt: Reptile-D-Learning for Robust and Efficient Control Under Parametric Uncertainty

    cs.RO 2026-06 unverdicted novelty 5.0

    Reptile-D-learning applies meta-learning to produce generalizable Lyapunov network initializations that enable rapid adaptation of stable controllers to unseen parametric variations in nonlinear systems.