Xue-Ping Guo
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Xue-Ping Guo has a total of 19 co-authors. Here are the top co-authors by number of shared publications:
- Long-Ze Tan
Key Laboratory of MEA (Ministry of Education) & Shanghai Key Laboratory of PMMP, School of Mathematical Sciences, East China Normal University, Shanghai, China
3 shared publications - Guoliang Chen
Department of Mathematics, East China Normal University, Shanghai, China
2 shared publications - Hong-Xiu Zhong
School of Mathematics and Data Science, Jiangnan University, Wuxi, People’s Republic of China
2 shared publications - Jiali Qiu
School of Mathematics and Statistics, Xi'an Jiaotong University, Shaanxi 710049, China
2 shared publications - Jing Wang
Shandong Computer Science Center, Jinan, Shandong, People’s Republic of China
2 shared publications Long-Ze Tan
University of Science and Technology of China
2 shared publications
Latest Publications
- Open Access
- 15 May 2026
- Published by Global Science Press in Journal of Computational Mathematics
- Vol. 44 (5) , 1330-1353
On the Adaptive Deterministic Block Coordinate Descent Method with Momentum for Solving Large Linear Least-Squares Problems
Inspired by Polyak’s heavy-ball method, this paper proposes an adaptive deterministic block coordinate descent method with momentum (mADBCD) for efficiently solving large-scale linear least-squares problems. The proposed method introduces a novel column selection criterion based on the Euclidean norm of the residual vector of the normal equation. In contrast to classical block coordinate descent methods, mADBCD does not require a fixed pre-partitioning of the column indices of the coefficient matrix and avoids the expensive computation of Moore-Penrose pseudoinverses of submatrices at each iteration. The method adaptively updates the block index set at each step, thereby improving both flexibility and scalability. When the coefficient matrix is of full column rank, we prove that mADBCD converges linearly to the unique solution of the least-squares problem. Numerical experiments are conducted to show that mADBCD outperforms several recent block coordinate descent methods in terms of iteration count and CPU time. In particular, when solving extremely sparse least-squares problems, mADBCD is the first block coordinate descent method reported to achieve CPU time nearly comparable to that of the classical least squares QR (LSQR) method [Paige and Saunders, ACM Trans. Math. Softw., 8 (1982)].
- 15 May 2026
- Published by Global Science Press in Journal of Computational Mathematics
- Vol. 44 (5) , 1330-1353
Open AccessOn the Adaptive Deterministic Block Coordinate Descent Method with Momentum for Solving Large Linear Least-Squares Problems
Inspired by Polyak’s heavy-ball method, this paper proposes an adaptive deterministic block coordinate descent method with momentum (mADBCD) for efficiently solving large-scale linear least-squares problems. The proposed method introduces a novel column selection criterion based on the Euclidean norm of the residual vector of the normal equation. In contrast to classical block coordinate descent methods, mADBCD does not require a fixed pre-partitioning of the column indices of the coefficient matrix and avoids the expensive computation of Moore-Penrose pseudoinverses of submatrices at each iteration. The method adaptively updates the block index set at each step, thereby improving both flexibility and scalability. When the coefficient matrix is of full column rank, we prove that mADBCD converges linearly to the unique solution of the least-squares problem. Numerical experiments are conducted to show that mADBCD outperforms several recent block coordinate descent methods in terms of iteration count and CPU time. In particular, when solving extremely sparse least-squares problems, mADBCD is the first block coordinate descent method reported to achieve CPU time nearly comparable to that of the classical least squares QR (LSQR) method [Paige and Saunders, ACM Trans. Math. Softw., 8 (1982)].
- Open Access
- 25 November 2025
- Published by MDPI AG in Mathematics
- Vol. 13 (23) , 3781
NRSAI: An Adaptive and Efficient Preconditioner for Large Sparse Linear Systems
This paper presents a new Residual-based Sparse Approximate Inverse algorithm, designed to be automatic, computationally efficient, and highly parallelizable. After briefly reviewing Frobenius-norm-based preconditioning techniques and identifying two key challenges in this class of methods, we introduce an improved approach that integrates adaptive strategies from the Power Sparse Approximate Inverse algorithm and incomplete LU factorization. The method leverages the Hamilton–Cayley theorem for effective sparsity pattern construction and employs LU decomposition for efficient implementation. Theoretical analysis establishes the convergence properties and computational features of the proposed algorithm. Numerical experiments using real-world data demonstrate that the proposed algorithm significantly outperforms established methods—including the Sparse Approximate Inverse, Power Sparse Approximate Inverse and Residual-based Sparse Approximate Inverse algorithms—with a Generalized Minimal Residual iterative solver, confirming its superior efficiency and practical applicability.
- 25 November 2025
- Published by MDPI AG in Mathematics
- Vol. 13 (23) , 3781
Open AccessNRSAI: An Adaptive and Efficient Preconditioner for Large Sparse Linear Systems
This paper presents a new Residual-based Sparse Approximate Inverse algorithm, designed to be automatic, computationally efficient, and highly parallelizable. After briefly reviewing Frobenius-norm-based preconditioning techniques and identifying two key challenges in this class of methods, we introduce an improved approach that integrates adaptive strategies from the Power Sparse Approximate Inverse algorithm and incomplete LU factorization. The method leverages the Hamilton–Cayley theorem for effective sparsity pattern construction and employs LU decomposition for efficient implementation. Theoretical analysis establishes the convergence properties and computational features of the proposed algorithm. Numerical experiments using real-world data demonstrate that the proposed algorithm significantly outperforms established methods—including the Sparse Approximate Inverse, Power Sparse Approximate Inverse and Residual-based Sparse Approximate Inverse algorithms—with a Generalized Minimal Residual iterative solver, confirming its superior efficiency and practical applicability.
- Open Access
- 14 July 2025
- Published by Springer Nature in Research Square
Alternating Anderson Acceleration-like Algorithms for Solving Large Sparse Linear Systems
The increasing scale of scientific and engineering computations presents significant challenges in efficiently solving large sparse linear systems. In this work, we develop a class of enhanced iterative solvers by integrating Anderson Acceleration with the Hermitian and skew-Hermitian splitting (HSS) framework. Specifically, we propose the alternating Anderson-accelerated HSS (AAHSS) method, which incorporates the weighted HSS iterations and periodically applies Anderson Acceleration to improve convergence. A spectral radius-based convergence analysis is conducted to theoretically support the proposed strategy. Building upon the AAHSS method, we further design its four algorithmic variants: the single-step Anderson-accelerated HSS (AASHSS) method, its preconditioned version (AASPHSS), a parameterized single-step variant (AAPSHSS), and the parameterized preconditioned single-step (AAPSPHSS) methods. Comprehensive numerical experiments on a range of benchmark problems demonstrate that all proposed methods significantly reduce iteration counts and improve computational efficiency compared to classical Krylov subspace methods, such as the alternating Anderson-accelerated Richardson (AAR), GMRES, BiCGStab, Quasi-Minimal Residual (QMR) and CG methods. Among them, the AAPSHSS method performs particularly well on nonsymmetric systems, while the AAPSPHSS method is especially effective for symmetric problems.
- 14 July 2025
- Published by Springer Nature in Research Square
Open AccessAlternating Anderson Acceleration-like Algorithms for Solving Large Sparse Linear Systems
The increasing scale of scientific and engineering computations presents significant challenges in efficiently solving large sparse linear systems. In this work, we develop a class of enhanced iterative solvers by integrating Anderson Acceleration with the Hermitian and skew-Hermitian splitting (HSS) framework. Specifically, we propose the alternating Anderson-accelerated HSS (AAHSS) method, which incorporates the weighted HSS iterations and periodically applies Anderson Acceleration to improve convergence. A spectral radius-based convergence analysis is conducted to theoretically support the proposed strategy. Building upon the AAHSS method, we further design its four algorithmic variants: the single-step Anderson-accelerated HSS (AASHSS) method, its preconditioned version (AASPHSS), a parameterized single-step variant (AAPSHSS), and the parameterized preconditioned single-step (AAPSPHSS) methods. Comprehensive numerical experiments on a range of benchmark problems demonstrate that all proposed methods significantly reduce iteration counts and improve computational efficiency compared to classical Krylov subspace methods, such as the alternating Anderson-accelerated Richardson (AAR), GMRES, BiCGStab, Quasi-Minimal Residual (QMR) and CG methods. Among them, the AAPSHSS method performs particularly well on nonsymmetric systems, while the AAPSPHSS method is especially effective for symmetric problems.
- Open Access
On the adaptive deterministic block coordinate descent methods with momentum for solving large linear least-squares problems
In this work, we first present an adaptive deterministic block coordinate descent method with momentum (mADBCD) to solve the linear least-squares problem, which is based on Polyak's heavy ball method and a new column selection criterion for a set of block-controlled indices defined by the Euclidean norm of the residual vector of the normal equation. The mADBCD method eliminates the need for pre-partitioning the column indexes of the coefficient matrix, and it also obviates the need to compute the Moore-Penrose pseudoinverse of a column sub-matrix at each iteration. Moreover, we demonstrate the adaptability and flexibility in the automatic selection and updating of the block control index set. When the coefficient matrix has full rank, the theoretical analysis of the mADBCD method indicates that it linearly converges towards the unique solution of the linear least-squares problem. Furthermore, by effectively integrating count sketch technology with the mADBCD method, we also propose a novel count sketch adaptive block coordinate descent method with momentum (CS-mADBCD) for solving highly overdetermined linear least-squares problems and analysis its convergence. Finally, numerical experiments illustrate the advantages of the proposed two methods in terms of both CPU times and iteration counts compared to recent block coordinate descent methods.
Open AccessOn the adaptive deterministic block coordinate descent methods with momentum for solving large linear least-squares problems
In this work, we first present an adaptive deterministic block coordinate descent method with momentum (mADBCD) to solve the linear least-squares problem, which is based on Polyak's heavy ball method and a new column selection criterion for a set of block-controlled indices defined by the Euclidean norm of the residual vector of the normal equation. The mADBCD method eliminates the need for pre-partitioning the column indexes of the coefficient matrix, and it also obviates the need to compute the Moore-Penrose pseudoinverse of a column sub-matrix at each iteration. Moreover, we demonstrate the adaptability and flexibility in the automatic selection and updating of the block control index set. When the coefficient matrix has full rank, the theoretical analysis of the mADBCD method indicates that it linearly converges towards the unique solution of the linear least-squares problem. Furthermore, by effectively integrating count sketch technology with the mADBCD method, we also propose a novel count sketch adaptive block coordinate descent method with momentum (CS-mADBCD) for solving highly overdetermined linear least-squares problems and analysis its convergence. Finally, numerical experiments illustrate the advantages of the proposed two methods in terms of both CPU times and iteration counts compared to recent block coordinate descent methods.
- 25 March 2024
- Published by Springer Nature in Communications on Applied Mathematics and Computation
- Vol. 7 (4) , 1580-1597
On Multi-step Greedy Kaczmarz Method for Solving Large Sparse Consistent Linear Systems
- 25 March 2024
- Published by Springer Nature in Communications on Applied Mathematics and Computation
- Vol. 7 (4) , 1580-1597
On Multi-step Greedy Kaczmarz Method for Solving Large Sparse Consistent Linear Systems
- Open Access
- 20 January 2023
- Published by Springer Nature in Research Square
On nonlinear greedy randomized Kaczmarz methods for solving large-scale systems of nonlinear equations
The greedy randomized Kaczmarz (GRK) method is an efficient method for solving large-scale systems of linear equations. Based on this basic idea, we propose a nonlinear greedy randomized Kacz-marz (NGRK) method for solving large-scale systems of nonlinear equations and prove its convergence. Then, to improve the computational efficiency of the NGRK method, we design the multi-step NGRK (MNGRK) method and the accelerated NGK (ANGK) method, respectively. Theoretical results show that the convergence speed of the MNGRK and ANGK methods is faster than that of the NGRK method. Finally, inspired by the max-distance coordinate descent (MDCD) method, we also establish the nonlinear max-distance Kaczmarz (NMDK) method and its accelerated version.
- 20 January 2023
- Published by Springer Nature in Research Square
Open AccessOn nonlinear greedy randomized Kaczmarz methods for solving large-scale systems of nonlinear equations
The greedy randomized Kaczmarz (GRK) method is an efficient method for solving large-scale systems of linear equations. Based on this basic idea, we propose a nonlinear greedy randomized Kacz-marz (NGRK) method for solving large-scale systems of nonlinear equations and prove its convergence. Then, to improve the computational efficiency of the NGRK method, we design the multi-step NGRK (MNGRK) method and the accelerated NGK (ANGK) method, respectively. Theoretical results show that the convergence speed of the MNGRK and ANGK methods is faster than that of the NGRK method. Finally, inspired by the max-distance coordinate descent (MDCD) method, we also establish the nonlinear max-distance Kaczmarz (NMDK) method and its accelerated version.
- 7 January 2023
- Published by Springer Nature in Computational and Applied Mathematics
- Vol. 42 (1) , 1-20
On multi-step greedy randomized coordinate descent method for solving large linear least-squares problems
- 7 January 2023
- Published by Springer Nature in Computational and Applied Mathematics
- Vol. 42 (1) , 1-20
On multi-step greedy randomized coordinate descent method for solving large linear least-squares problems
- Open Access
- 1 January 2019
- Published by American Institute of Mathematical Sciences (AIMS) in Numerical Algebra, Control & Optimization
- Vol. 9 (1) , 85-99
Semi-local convergence of the Newton-HSS method under the center Lipschitz condition
Newton-type methods have gained much attention in the past decades, especially for the semilocal convergence based on no information around the solution $x_*$ of the target nonlinear equation. For large sparse non-Hermitian positive definite systems of nonlinear equation, assuming that the nonlinear operator satisfies the center Lipschitz condition, which is wider than usual Lipschtiz condition and H$ö$lder continuous condition, we establish a new Newton-Kantorovich convergence theorem for the Newton-HSS method. Once the convergence criteria is satisfied, the iteration sequence $\{x_k\}_{k = 0}^∞$ generated by the Newton-HSS method is well defined, and converges to the solution $x_*$. Numerical results illustrate the effect.- 1 January 2019
- Published by American Institute of Mathematical Sciences (AIMS) in Numerical Algebra, Control & Optimization
- Vol. 9 (1) , 85-99
Open AccessSemi-local convergence of the Newton-HSS method under the center Lipschitz condition
Newton-type methods have gained much attention in the past decades, especially for the semilocal convergence based on no information around the solution $x_*$ of the target nonlinear equation. For large sparse non-Hermitian positive definite systems of nonlinear equation, assuming that the nonlinear operator satisfies the center Lipschitz condition, which is wider than usual Lipschtiz condition and H$ö$lder continuous condition, we establish a new Newton-Kantorovich convergence theorem for the Newton-HSS method. Once the convergence criteria is satisfied, the iteration sequence $\{x_k\}_{k = 0}^∞$ generated by the Newton-HSS method is well defined, and converges to the solution $x_*$. Numerical results illustrate the effect. - 5 February 2018
- Published by Springer Nature in Numerical Algorithms
- Vol. 79 (4) , 1049-1073
On the accelerated modified Newton-HSS method for systems of nonlinear equations
- 5 February 2018
- Published by Springer Nature in Numerical Algorithms
- Vol. 79 (4) , 1049-1073
On the accelerated modified Newton-HSS method for systems of nonlinear equations