Strong Fast Invertibility and Lyapunov Exponents for Linear Systems
Florian Noethen

TL;DR
This paper characterizes strongly fast invertible linear systems with bounded coefficients, linking their properties to Lyapunov exponents and providing convergence results for Lyapunov exponent computation methods.
Contribution
It offers a full characterization of regular strongly fast invertible systems and establishes conditions for Lyapunov exponent convergence even without regularity.
Findings
Strongly fast invertibility relates to Lyapunov exponent multiplicities.
Stability of Lyapunov exponents implies strong fast invertibility.
Convergence conditions for Benettin's algorithm stepsizes are derived.
Abstract
In 2019 Anthony Quas, Philippe Thieullen and Mohamed Zarrabi introduced the concept of strong fast invertibility for linear cocycles. It relates the growth of volumes between different initial times and, together with a condition on singular value gaps, yields the existence of a dominated splitting of the dynamics. The properties of this splitting largely coincide with those for systems with stable Lyapunov exponents. In this article, we take a closer look at strongly fast invertible systems with bounded coefficients. By linking the dimensions at which a system admits strong fast invertibility to the multiplicities of Lyapunov exponents, we are able to give a full characterization of regular strongly fast invertible systems similar to that of systems with stable Lyapunov exponents. In particular, we show that the stability of Lyapunov exponents implies strong fast invertibility (even…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsModel Reduction and Neural Networks · Quantum chaos and dynamical systems · Advanced Thermodynamics and Statistical Mechanics
