Numerical optimization of eigenvalues of Hermitian matrix functions

DOI10.1137/130933472MaRDI QIDQ2923367zbMATH OpenOpenAlexFDO

Authors Emre Mengi, E. Alper Yildirim, Mustafa Kiliç

Publication date 15 October 2014

Published in SIAM Journal on Matrix Analysis and Applications (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1109.2080

zbMATH Keywords

algorithmanalytic and Hermitian matrix-valued functiondistance to uncontrollabilityglobal convergenceglobal optimizationHermitian eigenvaluesnumerical examplesnumerical radiusperturbation of eigenvalues

Mathematics Subject Classification ID

Numerical computation of eigenvalues and eigenvectors of matrices (65F15) Numerical computation of matrix exponential and similar matrix functions (65F60) Nonconvex programming, global optimization (90C26)

Abstract: This work concerns the global minimization of a prescribed eigenvalue or a weighted sum of prescribed eigenvalues of a Hermitian matrix-valued function depending on its parameters analytically in a box. We describe how the analytical properties of eigenvalue functions can be put into use to derive piece-wise quadratic functions that underestimate the eigenvalue functions. These piece-wise quadratic under-estimators lead us to a global minimization algorithm, originally due to Breiman and Cutler. We prove the global convergence of the algorithm, and show that it can be effectively used for the minimization of extreme eigenvalues, e.g., the largest eigenvalue or the sum of the largest specified number of eigenvalues. This is particularly facilitated by the analytical formulas for the first derivatives of eigenvalues, as well as analytical lower bounds on the second derivatives that can be deduced for extreme eigenvalue functions. The applications that we have in mind also include the ${mH}_{i}nfty$-norm of a linear dynamical system, numerical radius, distance to uncontrollability and various other non-convex eigenvalue optimization problems, for which, generically, the eigenvalue function involved is simple at all points.

Review text

This paper presents an efficient algorithm based on the analytical properties of eigenvalues of an analytic and Hermitian matrix-valued function, which can be used to optimize any eigenvalue problem where lower bounds on the second derivatives of the eigenvalue function can be calculated analytically or numerically. Among these problems the authors include computation of quantities related to dynamical systems, minimization of the largest and maximization of the smallest eigenvalue, minimization of the sum of the \(j\) largest eigenvalues, computation of numerical radius and distance to uncontrollability. Numerical examples related to these applications are also discussed in detail and comparisons with other algorithms are provided, which show the effectiveness of the presented algorithm. Global convergence of the algorithm is proved. Remarkably, a MATLAB implementation of the algorithm and a user guide are available on the web page of the first author.

Recommendations

• scientific article; zbMATH DE number 4096706
• A Support Function Based Algorithm for Optimization with Eigenvalue Constraints
• On minimizing the largest eigenvalue of a symmetric matrix
• Large-Scale Optimization of Eigenvalues
• A quadratically convergent local algorithm on minimizing sums of the largest eigenvalues of a symmetric matrix

Cited in

(31)

• Approximation of stability radii for large-scale dissipative Hamiltonian systems
• Generalized derivatives of eigenvalues of a symmetric matrix
• Black-box learning of multigrid parameters
• Inheritance properties of Krylov subspace methods for continuous-time algebraic Riccati equations
• Krylov subspace methods for discrete-time algebraic Riccati equations
• An unconstrained global optimization framework for real symmetric eigenvalue problems
• Robust stability optimization for linear delay systems in a probabilistic framework
• Subspace acceleration for large-scale parameter-dependent Hermitian eigenproblems
• A Support Function Based Algorithm for Optimization with Eigenvalue Constraints
• A subspace method for large-scale eigenvalue optimization
• Large-Scale Optimization of Eigenvalues
• Subspace acceleration for the Crawford number and related eigenvalue optimization problems
• Large-scale computation of \(\mathcal{L}_\infty\)-norms by a greedy subspace method
• Large-scale and global maximization of the distance to instability
• Matrix polynomials with specified eigenvalues
• Derivatives of symplectic eigenvalues and a Lidskii type theorem
• Nonlinear eigenvector methods for convex minimization over the numerical range
• Generating eigenvalue bounds using optimization
• Approximate residual-minimizing shift parameters for the low-rank ADI iteration
• Nonsmooth optimization method for \(H_\infty\) output feedback control
• Differential equations for real-structured defectivity measures
• A subspace framework for \(\mathcal{H}_\infty \)-norm minimization
• Model Order Reduction in Contour Integral Methods for Parametric PDEs
• Model order reduction for delay systems by iterative interpolation
• Variational characterization and Rayleigh quotient iteration of 2D eigenvalue problem with applications
• Spectrally constrained optimization
• Large-scale minimization of the pseudospectral abscissa
• Minimization of the pseudospectral abscissa of a quadratic matrix polynomial
• Detecting large definite Hermitian matrix pairs by -subspace algorithms
• A subspace framework for \(\mathcal{L}_{\infty}\) model reduction
• An indefinite LOBPCG type of algorithm for detecting a definite Hermitian matrix pair

Describes a project that uses

Uses Software

• CTDSX
• Matlab

This page was built for publication: Numerical optimization of eigenvalues of Hermitian matrix functions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2923367)