4.3 Article

HANKEL MATRIX RANK MINIMIZATION WITH APPLICATIONS TO SYSTEM IDENTIFICATION AND REALIZATION

期刊

出版社

SIAM PUBLICATIONS
DOI: 10.1137/110853996

关键词

rank minimization; nuclear norm; Hankel matrix; first-order method; system identification; system realization

向作者/读者索取更多资源

We introduce a flexible optimization framework for nuclear norm minimization of matrices with linear structure, including Hankel, Toeplitz, and moment structures and catalog applications from diverse fields under this framework. We discuss various first-order methods for solving the resulting optimization problem, including alternating direction methods of multipliers, proximal point algorithms, and gradient projection methods. We perform computational experiments to compare these methods on system identification problems and system realization problems. For the system identification problem, the gradient projection method (accelerated by Nesterov's extrapolation techniques) and the proximal point algorithm usually outperform other first-order methods in terms of CPU time on both real and simulated data, for small and large regularization parameters, respectively, while for the system realization problem, the alternating direction method of multipliers, as applied to a certain primal reformulation, usually outperforms other first-order methods in terms of CPU time. We also study the convergence of the proximal alternating direction methods of multipliers used in this paper.

作者

我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。

评论

主要评分

4.3
评分不足

次要评分

新颖性
-
重要性
-
科学严谨性
-
评价这篇论文

推荐

Article Computer Science, Software Engineering

An efficient Hessian based algorithm for solving large-scale sparse group Lasso problems

Yangjing Zhang, Ning Zhang, Defeng Sun, Kim-Chuan Toh

MATHEMATICAL PROGRAMMING (2020)

Article Computer Science, Software Engineering

On the efficient computation of a generalized Jacobian of the projector over the Birkhoff polytope

Xudong Li, Defeng Sun, Kim-Chuan Toh

MATHEMATICAL PROGRAMMING (2020)

Article Computer Science, Software Engineering

SDPNAL plus : A Matlab software for semidefinite programming with bound constraints (version 1.0)

Defeng Sun, Kim-Chuan Toh, Yancheng Yuan, Xin-Yuan Zhao

OPTIMIZATION METHODS & SOFTWARE (2020)

Article Operations Research & Management Science

Inner approximating the completely positive cone via the cone of scaled diagonally dominant matrices

Joao Gouveia, Ting Kei Pong, Mina Saee

JOURNAL OF GLOBAL OPTIMIZATION (2020)

Article Operations Research & Management Science

A difference-of-convex approach for split feasibility with applications to matrix factorizations and outlier detection

Chen Chen, Ting Kei Pong, Lulin Tan, Liaoyuan Zeng

JOURNAL OF GLOBAL OPTIMIZATION (2020)

Article Operations Research & Management Science

A strictly contractive Peaceman-Rachford splitting method for the doubly nonnegative relaxation of the minimum cut problem

Xinxin Li, Ting Kei Pong, Hao Sun, Henry Wolkowicz

Summary: The paper introduces a new algorithm for solving the minimum cut problem, utilizing strengthened SDP and DNN relaxations to evaluate upper and lower bounds efficiently. The use of the Peaceman-Rachford splitting method accelerates the computation process, and empirical results demonstrate the efficiency and robustness of the proposed method on random datasets and vertex separator problems.

COMPUTATIONAL OPTIMIZATION AND APPLICATIONS (2021)

Article Computer Science, Theory & Methods

Kurdyka-Lojasiewicz Exponent via Inf-projection

Peiran Yu, Guoyin Li, Ting Kei Pong

Summary: The KL exponent is crucial in estimating the convergence rate of many first-order optimization methods, with a value of 1/2 indicating local linear convergence. It is generally difficult to estimate, but can be preserved through inf-projection. The KL exponent of important convex optimization models can be maintained at 1/2 under specific conditions, and for nonconvex models, can be derived from their majorant functions.

FOUNDATIONS OF COMPUTATIONAL MATHEMATICS (2022)

Article Operations Research & Management Science

ρ-regularization subproblems: strong duality and an eigensolver-based algorithm

Liaoyuan Zeng, Ting Kei Pong

Summary: This paper studies TR type methods and proposes a general regularized subproblem. It derives a strong duality theorem and necessary and sufficient optimality condition for the subproblem. An eigensolver-based algorithm is also proposed to solve the derived dual problem. Numerical results are presented to validate the effectiveness of the algorithm.

COMPUTATIONAL OPTIMIZATION AND APPLICATIONS (2022)

Article Operations Research & Management Science

Doubly iteratively reweighted algorithm for constrained compressed sensing models

Shuqin Sun, Ting Kei Pong

Summary: We propose a new algorithmic framework for constrained compressed sensing models with nonconvex sparsity-inducing regularizers and nonconvex loss functions. Our framework efficiently solves subproblems using well-developed solvers and introduces a new termination criterion for infeasible solutions. Numerical comparisons demonstrate that our approaches outperform existing methods in terms of recovery errors and speed, particularly for badly scaled measurement matrices.

COMPUTATIONAL OPTIMIZATION AND APPLICATIONS (2023)

Article Mathematics, Applied

CONVERGENCE RATE ANALYSIS OF A SEQUENTIAL CONVEX PROGRAMMING METHOD WITH LINE SEARCH FOR A CLASS OF CONSTRAINED DIFFERENCE-OF-CONVEX OPTIMIZATION PROBLEMS

Peiran Yu, Ting Kei Pong, Zhaosong Lu

Summary: This paper studies the sequential convex programming method with monotone line search (SCPls) for a class of difference-of-convex optimization problems with smooth inequality constraints. The convergence rate of the sequence generated by SCPls in nonconvex and convex settings under suitable Kurdyka-Lojasiewicz (KL) assumptions is analyzed. The paper also discusses how to deduce the KL exponent of the extended objective function from its Lagrangian in convex settings.

SIAM JOURNAL ON OPTIMIZATION (2021)

Article Mathematics, Applied

ANALYSIS AND ALGORITHMS FOR SOME COMPRESSED SENSING MODELS BASED ON L1/L2 MINIMIZATION

Liaoyuan Zeng, Peiran Yu, Ting Kei Pong

Summary: The paper explores the use of the l(1)/l(2) norm ratio in compressed sensing problems, proposes an algorithm for noise situations, and demonstrates through numerical experiments that the algorithm can recover the original sparse vectors with reasonable accuracy.

SIAM JOURNAL ON OPTIMIZATION (2021)

Article Mathematics, Applied

A HYBRID PENALTY METHOD FOR A CLASS OF OPTIMIZATION PROBLEMS WITH MULTIPLE RANK CONSTRAINTS

Tianxiang Liu, Ivan Markovsky, Ting Kei Pong, Akiko Takeda

SIAM JOURNAL ON MATRIX ANALYSIS AND APPLICATIONS (2020)

Article Mathematics, Applied

A SUBGRADIENT-BASED APPROACH FOR FINDING THE MAXIMUM FEASIBLE SUBSYSTEM WITH RESPECT TO A SET

Minglu Ye, Ting Kei Pong

SIAM JOURNAL ON OPTIMIZATION (2020)

Article Computer Science, Software Engineering

A multi-stage convex relaxation approach to noisy structured low-rank matrix recovery

Shujun Bi, Shaohua Pan, Defeng Sun

MATHEMATICAL PROGRAMMING COMPUTATION (2020)

Article Automation & Control Systems

Solving the OSCAR and SLOPE Models Using a Semismooth Newton-Based Augmented Lagrangian Method

Ziyan Luo, Defeng Sun, Kim-Chuan Toh, Naihua Xiu

JOURNAL OF MACHINE LEARNING RESEARCH (2019)

暂无数据