On Efficiently Computing the Eigenvalues of Limited-Memory Quasi-Newton Matrices

DOI10.1137/140997737zbMath1337.49048arXiv1411.7301OpenAlexW2964123951MaRDI QIDQ3195438

Jennifer B. Erway, Roummel F. Marcia

Publication date: 19 October 2015

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

Full work available at URL: https://arxiv.org/abs/1411.7301

zbMATH Keywords

large-scale optimizationeigenvaluesspectral decompositionQR decompositionlimited-memory quasi-Newton methodsquasi-Newton matrices

Mathematics Subject Classification ID

Numerical computation of eigenvalues and eigenvectors of matrices (65F15) Numerical mathematical programming methods (65K05) Large-scale problems in mathematical programming (90C06) Nonlinear programming (90C30) Newton-type methods (49M15) Methods of quasi-Newton type (90C53) Eigenvalues, singular values, and eigenvectors (15A18) Conditioning of matrices (15A12)

Related Items

A limited memory quasi-Newton trust-region method for box constrained optimization, On efficiently combining limited-memory and trust-region techniques, A new multipoint symmetric secant method with a dense initial matrix, Trust-region algorithms for training responses: machine learning methods using indefinite Hessian approximations, A limited-memory trust-region method for nonlinear optimization with many equality constraints, On solving large-scale limited-memory quasi-Newton equations, On solving L-SR1 trust-region subproblems, A new regularized quasi-Newton method for unconstrained optimization, Exact linesearch limited-memory quasi-Newton methods for minimizing a quadratic function, A dense initialization for limited-memory quasi-Newton methods

Uses Software

• L-BFGS
• mctoolbox
• GQTPAR

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• A practical method for solving large-scale TRS
• On the limited memory BFGS method for large scale optimization
• Linear and nonlinear programming.
• Convergence of quasi-Newton matrices generated by the symmetric rank one update
• Rank-one modification of the symmetric eigenproblem
• Representations of quasi-Newton matrices and their use in limited memory methods
• On efficiently combining limited-memory and trust-region techniques
• Numerical methods for general and structured eigenvalue problems.
• Computing a Trust Region Step
• Solving the quadratic trust-region subproblem in a low-memory BFGS framework
• Updating Quasi-Newton Matrices with Limited Storage
• Computing Optimal Locally Constrained Steps
• A Numerical Study of the Limited Memory BFGS Method and the Truncated-Newton Method for Large Scale Optimization
• Reorthogonalization and Stable Algorithms for Updating the Gram-Schmidt QR Factorization
• BFGS with Update Skipping and Varying Memory
• Numerical Optimization
• Trust Region Methods
• Methods for Modifying Matrix Factorizations
• Accuracy and Stability of Numerical Algorithms
• Quasi-newton algorithms generate identical points
• Some Modified Matrix Eigenvalue Problems