Data-Driven and Low-Rank Implementations of Balanced Singular Perturbation Approximation

DOI10.1137/23M155791XarXiv2303.05361WikidataQ128170530 ScholiaQ128170530MaRDI QIDQ6154963FDOQ6154963

Authors: Björn Liljegren-Sailer, I. V. Gosea 

Publication date: 16 February 2024

Published in: SIAM Journal on Scientific Computing (Search for Journal in Brave)

Abstract: Balanced Singular Perturbation Approximation (SPA) is a model order reduction method for linear time-invariant systems that guarantees asymptotic stability and for which there exists an a priori error bound. In that respect, it is similar to Balanced Truncation (BT). However, the reduced models obtained by SPA generally introduce better approximation in the lower frequency range and near steady-states, whereas BT is better suited for the higher frequency range. Even so, independently of the frequency range of interest, BT and its variants are more often applied in practice, since there exist more efficient algorithmic realizations thereof. In this paper, we aim at closing this practically-relevant gap for SPA. We propose two novel and efficient algorithms that are adapted for different settings. Firstly, we derive a low-rank implementation of SPA that is applicable in the large-scale setting. Secondly, a data-driven reinterpretation of the method is proposed that only requires input-output data, and thus, is realization-free. A main tool for our derivations is the reciprocal transformation, which induces a distinct view on implementing the method. While the reciprocal transformation and the characterization of SPA is not new, its significance for the practical algorithmic realization has been overlooked in the literature. Our proposed algorithms have well-established counterparts for BT, and as such, also a comparable computational complexity. The numerical performance of the two novel implementations is tested for several numerical benchmarks, and comparisons to their counterparts for BT as well as the existing implementations of SPA are made.

Full work available at URL: https://arxiv.org/abs/2303.05361

Recommendations

• Controller order reduction using singular perturbation approximation
• Singular perturbation approximation of balanced systems
• Parametric generalized singular perturbation approximation for model order reduction
• The generalised singular perturbation approximation for bounded real and positive real control systems
• Data-driven balancing of linear dynamical systems

zbMATH Keywords

linear systemsbalanced truncationNumerical quadraturenon-intrusive methodsdata-driven modelingbalanced singular perturbation approximationlow-rank implementationrealization-free

Mathematics Subject Classification ID

Large-scale systems (93A15) System structure simplification (93B11) Linear systems in control theory (93C05) Time-scale analysis and singular perturbations in control/observation systems (93C70) Asymptotic stability in control theory (93D20)

Cites Work

• Reduced basis methods for partial differential equations. An introduction
• Dynamic mode decomposition of numerical and experimental data
• Computational Methods for Linear Matrix Equations
• Dimension reduction of large-scale systems. Proceedings of a workshop, Oberwolfach, Germany, October 19--25, 2003.
• Approximation of Large-Scale Dynamical Systems
• Principal component analysis in linear systems: Controllability, observability, and model reduction
• Title not available (Why is that?)
• A survey of projection-based model reduction methods for parametric dynamical systems
• All optimal Hankel-norm approximations of linear multivariable systems and theirL,^∞-error bounds†
• A Cyclic Low-Rank Smith Method for Large Sparse Lyapunov Equations
• Low Rank Solution of Lyapunov Equations
• Direct methods for matrix Sylvester and Lyapunov equations
• MODEL REDUCTION FOR FLUIDS, USING BALANCED PROPER ORTHOGONAL DECOMPOSITION
• Efficient handling of complex shift parameters in the low-rank Cholesky factor ADI method
• Balanced Truncation Model Reduction of Large-Scale Dense Systems on Parallel Computers
• Synthesis of minimum roundoff noise fixed point digital filters
• Singular perturbation approximation of balanced systems
• Gramian-based model reduction for data-sparse systems
• Model Order Reduction by Balanced Proper Orthogonal Decomposition and by Rational Interpolation
• A framework for the solution of the generalized realization problem
• Singular perturbational model reduction of balanced systems
• Data-driven operator inference for nonintrusive projection-based model reduction
• 2 Balancing-related model reduction methods
• 3 Model order reduction based on moment-matching
• Model reduction and approximation. Theory and algorithms
• A quadrature framework for solving Lyapunov and Sylvester equations
• The generalised singular perturbation approximation for bounded real and positive real control systems
• Operator inference for non-intrusive model reduction of systems with non-polynomial nonlinear terms
• Applications
• Learning low-dimensional dynamical-system models from noisy frequency-response data with Loewner rational interpolation
• Data-driven balancing of linear dynamical systems

This page was built for publication: Data-Driven and Low-Rank Implementations of Balanced Singular Perturbation Approximation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6154963)