A note on approximating the nearest stable discrete-time descriptor systems with fixed rank

@article{Gillis2020ANO,
  title={A note on approximating the nearest stable discrete-time descriptor systems with fixed rank},
  author={Nicolas Gillis and Michael Karow and Punit Sharma},
  journal={Applied Numerical Mathematics},
  year={2020},
  volume={148},
  pages={131-139}
}

Figures and Tables from this paper

Solving matrix nearness problems via Hamiltonian systems, matrix factorization, and optimization
TLDR
These lectures notes, reviewing recent works addressing various problems of finding the nearest stable system to an unstable one, propose a parametrization that allows efficiently compute the neareststable system (for matrices and matrix pairs), allowing to compute the distance to stability.
Nearest Ω-stable matrix via Riemannian optimization
TLDR
The task of finding the nearest Ω-stable matrix to a certain matrix A, i.e., the nearest matrix with all its eigenvalues in a prescribed closed set Ω is described as an optimization problem on the Riemannian manifold of orthogonal (or unitary) matrices.
Learning Stable Models for Prediction and Control
TLDR
A formula is derived that describes the prediction error of Koopman representations for an arbitrary number of time steps, and which shows that stability constraints can improve the predictive accuracy over long horizons.
Learning Stable Koopman Embeddings
TLDR
A new data-driven method for learning stable models of nonlinear systems that lifts the original state space to a higher-dimensional linear manifold using Koopman embeddings and proves that every discrete-time nonlinear contracting model can be learnt in this framework.
Learning Memory-Efficient Stable Linear Dynamical Systems for Prediction and Control
TLDR
This work presents an optimization method that ensures stability at every step and iteratively improves the reconstruction error using gradient directions, and achieves an orders-of-magnitude improvement in reconstruction error and superior results in terms of control performance.
System Norm Regularization Methods for Koopman Operator Approximation
TLDR
DMD and DMD with control are reformulated as convex optimization problems with linear matrix inequality constraints and hard asymptotic stability constraints and system norm regularizers are considered as methods to improve the numerical conditioning of the approximate Koopman operator.
Memory-Efficient Learning of Stable Linear Dynamical Systems for Prediction and Control
TLDR
This work presents an optimization method that ensures stability at every step and iteratively improves the reconstruction error using gradient directions derived, and achieves an orders-of-magnitude improvement in reconstruction error and superior results in terms of control performance.
Nearest $\varOmega $-stable matrix via Riemannian optimization
TLDR
The task of finding the nearest $\Omega$-stable matrix to a certain matrix $A$ is described as an optimization problem on the Riemannian manifold of orthogonal (or unitary) matrices, and the resulting algorithm is remarkably fast on small-scale and medium-scale matrices.

References

SHOWING 1-10 OF 25 REFERENCES
Approximating the nearest stable discrete-time system
Nearest stable system using successive convex approximations
Stabilizing discrete-time linear systems
In this paper, we consider the problem of stabilizing discrete-time linear systems by computing a nearby stable matrix to an unstable one. To do so, we provide a new characterization for the set of
AN LMI APPROACH TOWARDS STABILIZATION OF DISCRETE-TIME DESCRIPTOR SYSTEMS
Abstract In this paper descriptor feedback control of high index or even non-regular linear discrete-time descriptor systems is considered. The approach is based on a linear matrix inequality (LMI)
On the stability radius of a generalized state-space system
Computing the nearest stable matrix pairs
TLDR
A reformulation of the nearest stable matrix pair problem with a simpler feasible set by introducing dissipative Hamiltonian (DH) matrix pairs that allows for a fast gradient method to obtain a nearby stable approximation of a given matrix pair.
Computing closest stable non-negative matrices
TLDR
It is shown, that for the distances measured in the max-norm, one can find exact solution of the corresponding nonconvex projection problems in polynomial time and this can be seen as an analogue of Kharitonov's theorem for non-negative matrices.
On Nonsingularity Verification of Uncertain Matrices Over a Quadratically Constrained Set
  • T. Zhou
  • Mathematics
    IEEE Transactions on Automatic Control
  • 2011
Nonsingularity of a matrix is investigated in this technical note that depends affinely on several uncertainty matrices whose real and imaginary parts are constrained by some quadratic forms. Based
Finding the Nearest Positive-Real System
TLDR
This paper shows that a system is extended strictly PR if and only if it can be written as a strict port-Hamiltonian system, and uses a fast gradient method to obtain a nearby PR system to a given non-PR system.
...
1
2
3
...