Biao Qu
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
Biao Qu has a total of 51 co-authors. Here are the top co-authors by number of shared publications:
- Meixia Li
Weifang University
4 shared publications - Jun Sun
School of Mathematics and Statistics, Linyi University, Linyi, People’s Republic of China
3 shared publications Binghua Liu
Institute of Operations Research
2 shared publicationsJing Zhao
School of Management, Qufu Normal University, Rizhao, Shandong 276826, China
2 shared publications- Lingchen Kong
Beijing Jiaotong University
2 shared publications Bin Xu
Institute of Optoelectronics, China
1 shared publication
Latest Publications
- 20 March 2026
- Published by World Scientific Pub Co Pte Ltd in International Journal of Foundations of Computer Science
- p. 1-16
A Proximal Gradient Algorithm with Dimensionality Reduction for a Class of Optimization Problems in Radiotherapy
The split feasibility problem (SFP) arises in many fields in the real world, such as signal processing, image reconstruction and medical care. In this paper, we extract a class of problems in radiotherapy — the sparse split feasibility problem (SSFP): to find some solutions of the split feasibility problem to satisfy some sparsity. It is meaningful to design an effective and practical solution algorithm to solve this problem. Firstly, the problem is transformed into a sparse constrained optimization problem with convex objective function, and then a dimensionally reduced proximal gradient algorithm is constructed to solve the problem. The convergence of the algorithm is also proved. Numerical examples are provided, demonstrating the results.
- 20 March 2026
- Published by World Scientific Pub Co Pte Ltd in International Journal of Foundations of Computer Science
- p. 1-16
A Proximal Gradient Algorithm with Dimensionality Reduction for a Class of Optimization Problems in Radiotherapy
The split feasibility problem (SFP) arises in many fields in the real world, such as signal processing, image reconstruction and medical care. In this paper, we extract a class of problems in radiotherapy — the sparse split feasibility problem (SSFP): to find some solutions of the split feasibility problem to satisfy some sparsity. It is meaningful to design an effective and practical solution algorithm to solve this problem. Firstly, the problem is transformed into a sparse constrained optimization problem with convex objective function, and then a dimensionally reduced proximal gradient algorithm is constructed to solve the problem. The convergence of the algorithm is also proved. Numerical examples are provided, demonstrating the results.
- 2 January 2026
- Published by Springer Nature
- p. 381-395
Inertial Bregman Proximal Alternating Linearized Minimization Method with Line Search for Non-convex Split Feasibility Problem
- 2 January 2026
- Published by Springer Nature
- p. 381-395
Inertial Bregman Proximal Alternating Linearized Minimization Method with Line Search for Non-convex Split Feasibility Problem
- 2 January 2026
- Published by Springer Nature
- p. 396-407
Convergence and Linear Convergence of Alternated Inertial Projection Algorithm for Split Feasibility Problems
- 2 January 2026
- Published by Springer Nature
- p. 396-407
Convergence and Linear Convergence of Alternated Inertial Projection Algorithm for Split Feasibility Problems
- 1 January 2026
- Published by Chongqing Normal University in Pacific Journal of Optimization
- p. 21
The CQ Algorithm with Feasible Inexact Projection for the Split Feasibility Problem
(Communicated by Fanwen Meng) The split feasibility problem (SFP) arises in many applications such as signal processing and image reconstruction. A widely used approach for solving the SFP is the $CQ$ algorithm: $ x^{k+1} = P_C\big(x^k + \gamma A^T(P_Q - I)Ax^k\big), \ k = 0,1,\dots $ where $C\subset\mathbb{R}^n$ and $Q\subset\mathbb{R}^m$ are nonempty closed convex sets, $A\in\mathbb{R}^{m\times n}$, $\gamma\in \left(0,\frac{2}{L}\right)$, and $L$ represents the maximum eigenvalue of $A^T A$. However, this update requires computing exact projections with high accuracy, and carrying out such exact computations can be very expensive. In this paper, we propose a $CQ$-type method that replaces the orthogonal projection onto $C$ with a feasible inexact projection whose accuracy is controllable. The scheme maintains feasibility of iterates, lowers per-iteration cost, and uses an Armijo-like line search to ensure sufficient decrease of the objective function. This approach thereby removes the requirement for a pre-defined step-size boundas well as the need to compute the maximum eigenvalue of $A^TA$. It is proved that the sequence generated by the algorithm is Quasi-Fej\'er convergent to the solution set when the SFP is consistent. Numerical experiments demonstrate the effectiveness of the algorithm and its superior computational efficiency.- 1 January 2026
- Published by Chongqing Normal University in Pacific Journal of Optimization
- p. 21
The CQ Algorithm with Feasible Inexact Projection for the Split Feasibility Problem
(Communicated by Fanwen Meng) The split feasibility problem (SFP) arises in many applications such as signal processing and image reconstruction. A widely used approach for solving the SFP is the $CQ$ algorithm: $ x^{k+1} = P_C\big(x^k + \gamma A^T(P_Q - I)Ax^k\big), \ k = 0,1,\dots $ where $C\subset\mathbb{R}^n$ and $Q\subset\mathbb{R}^m$ are nonempty closed convex sets, $A\in\mathbb{R}^{m\times n}$, $\gamma\in \left(0,\frac{2}{L}\right)$, and $L$ represents the maximum eigenvalue of $A^T A$. However, this update requires computing exact projections with high accuracy, and carrying out such exact computations can be very expensive. In this paper, we propose a $CQ$-type method that replaces the orthogonal projection onto $C$ with a feasible inexact projection whose accuracy is controllable. The scheme maintains feasibility of iterates, lowers per-iteration cost, and uses an Armijo-like line search to ensure sufficient decrease of the objective function. This approach thereby removes the requirement for a pre-defined step-size boundas well as the need to compute the maximum eigenvalue of $A^TA$. It is proved that the sequence generated by the algorithm is Quasi-Fej\'er convergent to the solution set when the SFP is consistent. Numerical experiments demonstrate the effectiveness of the algorithm and its superior computational efficiency.- 15 July 2025
- Published by World Scientific Pub Co Pte Ltd in Asia-Pacific Journal of Operational Research
- Vol. 43 (02)
CQ-Like Algorithm for Nonconvex Split Feasibility Problem with Multiple Output Sets
With the development of science and technology, many practical problems can be transformed into nonconvex problems. As far as we know, current research on nonconvex split feasibility problems, especially on nonconvex split feasibility problems with multiple output sets, remains relatively limited. To address this need, we propose a CQ-like algorithm to solve them. This new algorithm extends the applicability of the CQ algorithm from two sets to multiple sets, and optimizes the step size to reduce the number of iterations. The convergence and convergence rates of the proposed algorithm for semialgebraic sets are proven by using the Kurdyka–Łojasiewicz inequality. Additionally, we apply the algorithm to diverse signal and image recovery problems, and compare its effectiveness with other existing algorithms. Numerical experiments reveal that the proposed algorithm demonstrates superior efficiency and practicality.
- 15 July 2025
- Published by World Scientific Pub Co Pte Ltd in Asia-Pacific Journal of Operational Research
- Vol. 43 (02)
CQ-Like Algorithm for Nonconvex Split Feasibility Problem with Multiple Output Sets
With the development of science and technology, many practical problems can be transformed into nonconvex problems. As far as we know, current research on nonconvex split feasibility problems, especially on nonconvex split feasibility problems with multiple output sets, remains relatively limited. To address this need, we propose a CQ-like algorithm to solve them. This new algorithm extends the applicability of the CQ algorithm from two sets to multiple sets, and optimizes the step size to reduce the number of iterations. The convergence and convergence rates of the proposed algorithm for semialgebraic sets are proven by using the Kurdyka–Łojasiewicz inequality. Additionally, we apply the algorithm to diverse signal and image recovery problems, and compare its effectiveness with other existing algorithms. Numerical experiments reveal that the proposed algorithm demonstrates superior efficiency and practicality.
- 24 June 2025
- Published by World Scientific Pub Co Pte Ltd in Asia-Pacific Journal of Operational Research
- Vol. 43 (02)
Inertial Forward—Reflected—Backward Splitting Method with Linesearch for Nonconvex Split Feasibility Problem
This paper studies the nonconvex split feasibility problem (NCSFP) which can be applied widely in some practical problems, such as compressed sensing, intensity–modulated radiation therapy, matrix factorization, outlier detection problem, and sudoku puzzle. The NCSFP is transformed into an unconstrained optimization problem whose objective function is nonconvex and nonsmooth. Then, we first adopt the framework of Malitsky—Tam forward—reflected—backward splitting method, together with inertial strategy to accelerate the convergence and a linesearch procedure to avoid computing the Lipschitz constant of some related mapping simultaneously. By assuming the semi–algebraic sets, we demonstrate that each bounded sequence generated by inertial forward—reflected—backward splitting method with a linesearch globally converges to a critical point and obtain the linear convergence rates. Numerical experiments validate rationality and effectiveness of the proposed algorithm.
- 24 June 2025
- Published by World Scientific Pub Co Pte Ltd in Asia-Pacific Journal of Operational Research
- Vol. 43 (02)
Inertial Forward—Reflected—Backward Splitting Method with Linesearch for Nonconvex Split Feasibility Problem
This paper studies the nonconvex split feasibility problem (NCSFP) which can be applied widely in some practical problems, such as compressed sensing, intensity–modulated radiation therapy, matrix factorization, outlier detection problem, and sudoku puzzle. The NCSFP is transformed into an unconstrained optimization problem whose objective function is nonconvex and nonsmooth. Then, we first adopt the framework of Malitsky—Tam forward—reflected—backward splitting method, together with inertial strategy to accelerate the convergence and a linesearch procedure to avoid computing the Lipschitz constant of some related mapping simultaneously. By assuming the semi–algebraic sets, we demonstrate that each bounded sequence generated by inertial forward—reflected—backward splitting method with a linesearch globally converges to a critical point and obtain the linear convergence rates. Numerical experiments validate rationality and effectiveness of the proposed algorithm.
- 1 January 2025
- Published by American Institute of Mathematical Sciences (AIMS) in Journal of Industrial and Management Optimization
- Vol. 21 (10) , 6190-6218
Inertial alternating projection gradient algorithm for nonconvex split feasibility problem with multiple output sets
In this paper, we investigate the split feasibility problem with multiple output sets, where the sets can be nonconvex. To address this problem, we propose an inertial alternating projection gradient algorithm, which is an improvement over the existing projection algorithm. The new algorithm has a broader applicability and incorporates inertial terms to accelerate convergence. Next, we prove its convergence for semi-algebraic sets by using the Kurdyka-Łojasiewicz inequality, and subsequently establish the convergence rates of the proposed algorithm. Furthermore, we apply the proposed algorithm to signal and image recovery problems, comparing their performance with other algorithms. Our numerical results show that the proposed algorithm offers improved efficiency compared to existing methods in most cases.
- 1 January 2025
- Published by American Institute of Mathematical Sciences (AIMS) in Journal of Industrial and Management Optimization
- Vol. 21 (10) , 6190-6218
Inertial alternating projection gradient algorithm for nonconvex split feasibility problem with multiple output sets
In this paper, we investigate the split feasibility problem with multiple output sets, where the sets can be nonconvex. To address this problem, we propose an inertial alternating projection gradient algorithm, which is an improvement over the existing projection algorithm. The new algorithm has a broader applicability and incorporates inertial terms to accelerate convergence. Next, we prove its convergence for semi-algebraic sets by using the Kurdyka-Łojasiewicz inequality, and subsequently establish the convergence rates of the proposed algorithm. Furthermore, we apply the proposed algorithm to signal and image recovery problems, comparing their performance with other algorithms. Our numerical results show that the proposed algorithm offers improved efficiency compared to existing methods in most cases.
- 1 January 2025
- Published by Chongqing Normal University in Pacific Journal of Optimization
- p. 33
KM-CQ-like Alternating Inertial Algorithm for the Split Feasibility Problem
(Communicated by Fanwen Meng) The split feasibility problem (SFP) has many important and wide applications, such as radiotherapy, image reconstruction, and signal processing. Let $C$ and $Q$ be nonempty closed convex sets in $\Re^{n}$ and $\Re^{m}$, respectively, and $A$ an $m \times n$ real matrix, this kind of problem is about finding $x\in C, ~\mathrm{s.t.}~Ax\in Q,$ (1) if such $x$ exists. Recent researches on the SFP have been deeply integrated with inertial acceleration method, and some scholars have discovered that these techniques can speed up the process of first-order optimization algorithms. However, the most of them limit the range of inertial factors and cause the sequence $\|x^{k}-z\|(z\in S)$ no longer monotonically non-increasing. The alternating inertial method, expressed as $w^{k}=\begin{cases}x^{k},&\text{if ~}k\text{~is even},\\ x^{k}+\theta _{k}( x^{k}-x^{k-1}) ,&{\text{if ~}k\text{~is odd}},\end{cases}$ (2) is a appropriate improvement strategy, which retains the acceleration effect of inertial method while alleviating the degree of oscillation. In this paper, we propose three KM-CQ-like alternating inertial algorithms that all expand the range of inertial factors. We use projections onto general closed convex sets in the first version, which exploring the acceleration effects of alternating inertial technique preliminarily. The second version simplifies the computation of procedure by relaxing projections onto half-spaces. The last version further optimizes the previous methods by incorporating double projection technique. Our specific contributions in this paper are summarized as follows. (i) The proposed three KM-CQ-like algorithms incorporate alternating inertial steps allowing them to improve the convergence of the algorithms without inertial steps, which improve the range of values of inertial factors and simplify the form of limiting condition. (ii) The second algorithm also add relaxation effects that allows it to accelerate the convergence of the first algorithm and simplify the calculation of projections. On the other hand, we can see it is faster in Example 4.2. (iii) The last algorithm use double projections that are generated by half-spaces in the previous iteration and the current iteration. Similar to the expected conclusions, it converges faster than the second algorithm in Example 4.2. (iv) The convergence of the iterative sequences generated by the proposed algorithms is established, with the monotonicity of sequence $\|{x}^{2k}-z\|(z\in S)$, which alleviates the oscillation of the iterative points. (v) The performance and advantages of the algorithms proposed in this paper are confirmed by two applications in a simple the SFP and signal processing. On the other hand, the increase of dimension will make the advantages of our algorithms more obvious. Under simple parameter settings, these algorithms restore the monotonicity of $\|x^{2k}-z\|$ and achieve the convergence of the algorithm. Experiments demonstrate the feasibility of each algorithm in practical applications and the effectiveness of the acceleration procedure.- 1 January 2025
- Published by Chongqing Normal University in Pacific Journal of Optimization
- p. 33
KM-CQ-like Alternating Inertial Algorithm for the Split Feasibility Problem
(Communicated by Fanwen Meng) The split feasibility problem (SFP) has many important and wide applications, such as radiotherapy, image reconstruction, and signal processing. Let $C$ and $Q$ be nonempty closed convex sets in $\Re^{n}$ and $\Re^{m}$, respectively, and $A$ an $m \times n$ real matrix, this kind of problem is about finding $x\in C, ~\mathrm{s.t.}~Ax\in Q,$ (1) if such $x$ exists. Recent researches on the SFP have been deeply integrated with inertial acceleration method, and some scholars have discovered that these techniques can speed up the process of first-order optimization algorithms. However, the most of them limit the range of inertial factors and cause the sequence $\|x^{k}-z\|(z\in S)$ no longer monotonically non-increasing. The alternating inertial method, expressed as $w^{k}=\begin{cases}x^{k},&\text{if ~}k\text{~is even},\\ x^{k}+\theta _{k}( x^{k}-x^{k-1}) ,&{\text{if ~}k\text{~is odd}},\end{cases}$ (2) is a appropriate improvement strategy, which retains the acceleration effect of inertial method while alleviating the degree of oscillation. In this paper, we propose three KM-CQ-like alternating inertial algorithms that all expand the range of inertial factors. We use projections onto general closed convex sets in the first version, which exploring the acceleration effects of alternating inertial technique preliminarily. The second version simplifies the computation of procedure by relaxing projections onto half-spaces. The last version further optimizes the previous methods by incorporating double projection technique. Our specific contributions in this paper are summarized as follows. (i) The proposed three KM-CQ-like algorithms incorporate alternating inertial steps allowing them to improve the convergence of the algorithms without inertial steps, which improve the range of values of inertial factors and simplify the form of limiting condition. (ii) The second algorithm also add relaxation effects that allows it to accelerate the convergence of the first algorithm and simplify the calculation of projections. On the other hand, we can see it is faster in Example 4.2. (iii) The last algorithm use double projections that are generated by half-spaces in the previous iteration and the current iteration. Similar to the expected conclusions, it converges faster than the second algorithm in Example 4.2. (iv) The convergence of the iterative sequences generated by the proposed algorithms is established, with the monotonicity of sequence $\|{x}^{2k}-z\|(z\in S)$, which alleviates the oscillation of the iterative points. (v) The performance and advantages of the algorithms proposed in this paper are confirmed by two applications in a simple the SFP and signal processing. On the other hand, the increase of dimension will make the advantages of our algorithms more obvious. Under simple parameter settings, these algorithms restore the monotonicity of $\|x^{2k}-z\|$ and achieve the convergence of the algorithm. Experiments demonstrate the feasibility of each algorithm in practical applications and the effectiveness of the acceleration procedure.- 1 January 2025
- Published by American Institute of Mathematical Sciences (AIMS) in Mathematical Foundations of Computing
- Vol. 8 (3) , 419-432
Heavy-ball-based relaxed optimal s-thresholding algorithms for solving compressed sensing problem
This paper focuses on the problem of compressed sensing. By merging heavy-ball acceleration method which is a multi-step extension of the traditional gradient descent method and a new thresholding technique – optimal s-thresholding, we propose two algorithms to solve the compressed sensing problem, which are heavy-ball-based relaxed optimal s-thresholding algorithm (HBROT) and heavy-ball-based relaxed optimal s-thresholding pursuit algorithm (HBROTP). Moreover, we prove that all the above algorithms can accurately recover a s-sparse signal if the restricted isometry constant of the measurement matrix satisfies some conditions.
- 1 January 2025
- Published by American Institute of Mathematical Sciences (AIMS) in Mathematical Foundations of Computing
- Vol. 8 (3) , 419-432
Heavy-ball-based relaxed optimal s-thresholding algorithms for solving compressed sensing problem
This paper focuses on the problem of compressed sensing. By merging heavy-ball acceleration method which is a multi-step extension of the traditional gradient descent method and a new thresholding technique – optimal s-thresholding, we propose two algorithms to solve the compressed sensing problem, which are heavy-ball-based relaxed optimal s-thresholding algorithm (HBROT) and heavy-ball-based relaxed optimal s-thresholding pursuit algorithm (HBROTP). Moreover, we prove that all the above algorithms can accurately recover a s-sparse signal if the restricted isometry constant of the measurement matrix satisfies some conditions.
- 3 August 2023
- Published by International Society for Magnetic Resonance in Medicine (ISMRM)
XCloud-pFISTA: A Medical Intelligence Cloud for Accelerated MRI Reconstruction
Machine learning and artificial intelligence have shown remarkable performance in accelerated magnetic resonance imaging (MRI). Cloud computing technologies have great advantages in building an easily accessible platform to deploy advanced algorithms. In this work, we develop a high-performance medical intelligence cloud computing platform (XCloud-pFISTA) to reconstruct MRI images from undersampled k-space data. Two state-of-the-art approaches of the Projected Fast Iterative Soft-Thresholding Algorithm (pFISTA) family have been successfully implemented on the cloud. This work can be considered as a good example of cloud-based medical image reconstruction and may benefit the future development of integrated reconstruction and online diagnosis system.
- 3 August 2023
- Published by International Society for Magnetic Resonance in Medicine (ISMRM)
XCloud-pFISTA: A Medical Intelligence Cloud for Accelerated MRI Reconstruction
Machine learning and artificial intelligence have shown remarkable performance in accelerated magnetic resonance imaging (MRI). Cloud computing technologies have great advantages in building an easily accessible platform to deploy advanced algorithms. In this work, we develop a high-performance medical intelligence cloud computing platform (XCloud-pFISTA) to reconstruct MRI images from undersampled k-space data. Two state-of-the-art approaches of the Projected Fast Iterative Soft-Thresholding Algorithm (pFISTA) family have been successfully implemented on the cloud. This work can be considered as a good example of cloud-based medical image reconstruction and may benefit the future development of integrated reconstruction and online diagnosis system.