Liping Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Liping Zhang has a total of 57 co-authors. Here are the top co-authors by number of shared publications:
- Defeng Sun
Department of Applied Mathematics, The Hong Kong Polytechnic University, Hong Kong, China
4 shared publications - Yongjun Chen
Department of Mathematical Sciences, Tsinghua University, Beijing, China
4 shared publications Zhenyu Ming
Theory Lab Central Research Institute 2012
4 shared publications- Yan Li
Department of Mathematical Sciences, Tsinghua University, Beijing, 100084, China
3 shared publications - Liqun Qi
Department of Applied Mathematics, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong
2 shared publications - Xingbang Cui
School of Mathematics and Statistics, Shaanxi Normal University, Xi’an, China
2 shared publications
Latest Publications
- 30 March 2026
- Published by Elsevier in Optics & Laser Technology
- Vol. 201
Shift-window meets dual attention: A multi-model architecture for specular highlight removal
- 30 March 2026
- Published by Elsevier in Optics & Laser Technology
- Vol. 201
Shift-window meets dual attention: A multi-model architecture for specular highlight removal
- Open Access
- 23 March 2026
- Published by Elsevier in SSRN Electronic Journal
Efficient Solver via Iterative Reweighted Regularization ADMM for Inverse Lithography Problem
In advanced semiconductor manufacturing, optical lithography frequently encounters image distortions due to mask diffraction, which represents a critical bottleneck to achieving higher resolution. To mitigate this, the lithography process is conventionally modeled as an inverse problem for mask correction. However, this inverse formulation is inherently ill-posed, necessitating the incorporation of appropriate regularization. In this paper, we propose a novel iterative reweighted Tikhonov framework that retains the edge-preserving capability while significantly enhancing numerical efficiency. We reformulate the inverse problem using a pixel-wise weighted Tikhonov penalty and develop an efficient alternating direction method of multipliers with iteratively reweighted regularization (IRTR-ADMM) to solve this new model. The IRTR-ADMM transforms the original problem into an iterative procedure that alternately solves three tractable subproblems, where the regularization weights are adaptively updated based on the current estimate. Furthermore, by constructing a monotonic Lyapunov function, we establish the convergence properties of IRTR-ADMM under mild conditions. Extensive experiments demonstrate that IRTR-ADMM achieves high-quality mask optimization with a short runtime, effectively reducing lithographic errors and maintaining both fine structural details and sharp transitions in the printed images.
- 23 March 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessEfficient Solver via Iterative Reweighted Regularization ADMM for Inverse Lithography Problem
In advanced semiconductor manufacturing, optical lithography frequently encounters image distortions due to mask diffraction, which represents a critical bottleneck to achieving higher resolution. To mitigate this, the lithography process is conventionally modeled as an inverse problem for mask correction. However, this inverse formulation is inherently ill-posed, necessitating the incorporation of appropriate regularization. In this paper, we propose a novel iterative reweighted Tikhonov framework that retains the edge-preserving capability while significantly enhancing numerical efficiency. We reformulate the inverse problem using a pixel-wise weighted Tikhonov penalty and develop an efficient alternating direction method of multipliers with iteratively reweighted regularization (IRTR-ADMM) to solve this new model. The IRTR-ADMM transforms the original problem into an iterative procedure that alternately solves three tractable subproblems, where the regularization weights are adaptively updated based on the current estimate. Furthermore, by constructing a monotonic Lyapunov function, we establish the convergence properties of IRTR-ADMM under mild conditions. Extensive experiments demonstrate that IRTR-ADMM achieves high-quality mask optimization with a short runtime, effectively reducing lithographic errors and maintaining both fine structural details and sharp transitions in the printed images.
- 27 January 2026
- Published by Springer Nature in Computational Optimization and Applications
- Vol. 93 (3) , 1107-1143
New restricted isometry properties of constrained and unconstrained $L_{*-F}$ minimization for low-rank approximation- 27 January 2026
- Published by Springer Nature in Computational Optimization and Applications
- Vol. 93 (3) , 1107-1143
New restricted isometry properties of constrained and unconstrained $L_{*-F}$ minimization for low-rank approximation- 18 December 2025
- Published by Springer Nature in Computational and Applied Mathematics
- Vol. 45 (4) , 130
Improved power methods for computing eigenvalues of dual quaternion Hermitian matrices
- 18 December 2025
- Published by Springer Nature in Computational and Applied Mathematics
- Vol. 45 (4) , 130
Improved power methods for computing eigenvalues of dual quaternion Hermitian matrices
- Open Access
Exchange Policy Optimization Algorithm for Semi-Infinite Safe Reinforcement Learning
Safe reinforcement learning (safe RL) aims to respect safety requirements while optimizing long-term performance. In many practical applications, however, the problem involves an infinite number of constraints, known as semi-infinite safe RL (SI-safe RL). Such constraints typically appear when safety conditions must be enforced across an entire continuous parameter space, such as ensuring adequate resource distribution at every spatial location. In this paper, we propose exchange policy optimization (EPO), an algorithmic framework that achieves optimal policy performance and deterministic bounded safety. EPO works by iteratively solving safe RL subproblems with finite constraint sets and adaptively adjusting the active set through constraint expansion and deletion. At each iteration, constraints with violations exceeding the predefined tolerance are added to refine the policy, while those with zero Lagrange multipliers are removed after the policy update. This exchange rule prevents uncontrolled growth of the working set and supports effective policy training. Our theoretical analysis demonstrates that, under mild assumptions, strategies trained via EPO achieve performance comparable to optimal solutions with global constraint violations strictly remaining within a prescribed bound.
Open AccessExchange Policy Optimization Algorithm for Semi-Infinite Safe Reinforcement Learning
Safe reinforcement learning (safe RL) aims to respect safety requirements while optimizing long-term performance. In many practical applications, however, the problem involves an infinite number of constraints, known as semi-infinite safe RL (SI-safe RL). Such constraints typically appear when safety conditions must be enforced across an entire continuous parameter space, such as ensuring adequate resource distribution at every spatial location. In this paper, we propose exchange policy optimization (EPO), an algorithmic framework that achieves optimal policy performance and deterministic bounded safety. EPO works by iteratively solving safe RL subproblems with finite constraint sets and adaptively adjusting the active set through constraint expansion and deletion. At each iteration, constraints with violations exceeding the predefined tolerance are added to refine the policy, while those with zero Lagrange multipliers are removed after the policy update. This exchange rule prevents uncontrolled growth of the working set and supports effective policy training. Our theoretical analysis demonstrates that, under mild assumptions, strategies trained via EPO achieve performance comparable to optimal solutions with global constraint violations strictly remaining within a prescribed bound.
- 25 October 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 299
High-dimensional low-rank three-way tensor autoregressive time series predictor
- 25 October 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 299
High-dimensional low-rank three-way tensor autoregressive time series predictor
- Open Access
On Error Bounds for Rank-Constrained Affine Matrix Sets
Rank-constrained matrix problems appear frequently across science and engineering. The convergence analysis of iterative algorithms developed for these problems often hinges on local error bounds, which correlate the distance to the feasible set with a measure of how much the constraints are violated. Foundational results in semi-algebraic geometry guarantee that such bounds exist, yet the associated exponents are generally not explicitly determined. This paper establishes a local Hölderian error bound with an explicit exponent for the canonical rank-constrained affine feasibility set. This paper proves that, on any compact set, the distance to the feasible set is bounded by a power of a natural residual function capturing violations in both the rank and affine constraints. The exponent in this bound is given explicitly in terms of the problem's dimensions. This provides a fundamental quantitative result on the geometry of the solution set, paving the way for the convergence analysis of a broad class of numerical methods.
Open AccessOn Error Bounds for Rank-Constrained Affine Matrix Sets
Rank-constrained matrix problems appear frequently across science and engineering. The convergence analysis of iterative algorithms developed for these problems often hinges on local error bounds, which correlate the distance to the feasible set with a measure of how much the constraints are violated. Foundational results in semi-algebraic geometry guarantee that such bounds exist, yet the associated exponents are generally not explicitly determined. This paper establishes a local Hölderian error bound with an explicit exponent for the canonical rank-constrained affine feasibility set. This paper proves that, on any compact set, the distance to the feasible set is bounded by a power of a natural residual function capturing violations in both the rank and affine constraints. The exponent in this bound is given explicitly in terms of the problem's dimensions. This provides a fundamental quantitative result on the geometry of the solution set, paving the way for the convergence analysis of a broad class of numerical methods.
- Open Access
- 2 June 2025
- Published by Springer Nature in Mathematical Programming
- Vol. 215 (1-2) , 637-668
Characterizations of the Aubin property of the solution mapping for nonlinear semidefinite programming
In this paper, we study the Aubin property of the Karush-Kuhn-Tucker solution mapping for the nonlinear semidefinite programming (NLSDP) problem at a locally optimal solution. In the literature, it is known that the Aubin property implies the constraint nondegeneracy by Fusek (SIAM J. Optim. 23:1041–1061, 2013) and the second-order sufficient condition by Ding et al. (SIAM J. Optim. 27:67–90, 2017). Based on the Mordukhovich criterion, here we further prove that the strong second-order sufficient condition is also necessary for the Aubin property to hold. Consequently, several equivalent conditions including the strong regularity are established for NLSDP’s Aubin property. Together with the recent progress made by Chen et al. (SIAM J. Optim. 35:712–738, 2025) on the equivalence between the Aubin property and the strong regularity for nonlinear second-order cone programming, this paper constitutes a significant step forward in characterizing the Aubin property for general non-polyhedral $$C^2$$ -cone reducible constrained optimization problems.- 2 June 2025
- Published by Springer Nature in Mathematical Programming
- Vol. 215 (1-2) , 637-668
Open AccessCharacterizations of the Aubin property of the solution mapping for nonlinear semidefinite programming
In this paper, we study the Aubin property of the Karush-Kuhn-Tucker solution mapping for the nonlinear semidefinite programming (NLSDP) problem at a locally optimal solution. In the literature, it is known that the Aubin property implies the constraint nondegeneracy by Fusek (SIAM J. Optim. 23:1041–1061, 2013) and the second-order sufficient condition by Ding et al. (SIAM J. Optim. 27:67–90, 2017). Based on the Mordukhovich criterion, here we further prove that the strong second-order sufficient condition is also necessary for the Aubin property to hold. Consequently, several equivalent conditions including the strong regularity are established for NLSDP’s Aubin property. Together with the recent progress made by Chen et al. (SIAM J. Optim. 35:712–738, 2025) on the equivalence between the Aubin property and the strong regularity for nonlinear second-order cone programming, this paper constitutes a significant step forward in characterizing the Aubin property for general non-polyhedral $$C^2$$ -cone reducible constrained optimization problems. - Open Access
Improved power methods for computing eigenvalues of dual quaternion Hermitian matrices
This paper investigates the eigenvalue computation problem of the dual quaternion Hermitian matrix closely related to multi-agent group control. Recently, power method was proposed by Cui and Qi in Journal of Scientific Computing, 100 (2024) to solve such problem. Recognizing that the convergence rate of power method is slow due to its dependence on the eigenvalue distribution, we propose two improved versions of power method based on dual complex adjoint matrices and Aitken extrapolation, named DCAM-PM and ADCAM-PM. They achieve notable efficiency improvements and demonstrate significantly faster convergence. However, power method may be invalid for dual quaternion Hermitian matrices with eigenvalues having identical standard parts but distinct dual parts. To overcome this disadvantage, utilizing the eigen-decomposition properties of dual complex adjoint matrix, we propose a novel algorithm EDDCAM-EA which surpasses the power method in both accuracy and speed. Application to eigenvalue computations of dual quaternion Hermitian matrices in multi-agent formation control and numerical experiments highlight the remarkable accuracy and speed of our proposed algorithms.
Open AccessImproved power methods for computing eigenvalues of dual quaternion Hermitian matrices
This paper investigates the eigenvalue computation problem of the dual quaternion Hermitian matrix closely related to multi-agent group control. Recently, power method was proposed by Cui and Qi in Journal of Scientific Computing, 100 (2024) to solve such problem. Recognizing that the convergence rate of power method is slow due to its dependence on the eigenvalue distribution, we propose two improved versions of power method based on dual complex adjoint matrices and Aitken extrapolation, named DCAM-PM and ADCAM-PM. They achieve notable efficiency improvements and demonstrate significantly faster convergence. However, power method may be invalid for dual quaternion Hermitian matrices with eigenvalues having identical standard parts but distinct dual parts. To overcome this disadvantage, utilizing the eigen-decomposition properties of dual complex adjoint matrix, we propose a novel algorithm EDDCAM-EA which surpasses the power method in both accuracy and speed. Application to eigenvalue computations of dual quaternion Hermitian matrices in multi-agent formation control and numerical experiments highlight the remarkable accuracy and speed of our proposed algorithms.
- 16 May 2025
- Published by Springer Nature in Journal of Scientific Computing
- Vol. 104 (1) , 1-24
Dual Complex Adjoint Matrix and Applications to Hand-Eye Calibration and Multi-Agent Formation Control
- 16 May 2025
- Published by Springer Nature in Journal of Scientific Computing
- Vol. 104 (1) , 1-24
Dual Complex Adjoint Matrix and Applications to Hand-Eye Calibration and Multi-Agent Formation Control