Yanqin Bai
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yanqin Bai has a total of 95 co-authors. Here are the top co-authors by number of shared publications:
- Guoqiang Wang
School of Mathematics, Physics and Statistics, Shanghai University of Engineering Science, Shanghai, China
9 shared publications - Changjun Yu
Department of Mathematics, Shanghai University, Shanghai, China
8 shared publications - Qian Li
School of Mathematics, Physics and Statistics, Shanghai University of Engineering Science, Shanghai 201620, P. R. China
5 shared publications - Chuan-Hao Guo
Department of Mathematics Shanghai University 200444 Shanghai China
4 shared publications - Hai Xiang Lin
Delft Institute of Applied Mathematics Delft University of Technology Delft the Netherlands
4 shared publications - Kok Lay Teo
School of Mathematical Sciences, Sunway University, Bandar Sunway, Selangor Darul Ehsan, Malaysia
4 shared publications
Latest Publications
- 1 January 2025
- Published by American Institute of Mathematical Sciences (AIMS) in Journal of Industrial and Management Optimization
- Vol. 21 (7) , 4712-4729
Smoothed hybrid $ l_p $-$ l_2 $ model for sparse optimizationIn this paper, we propose a new regularization term by using a hybrid function of the smoothed sub-one $ l_p $ quasi-norm and $ l_2 $ norm. The new regularization term avoids the non-differentiability of the sub-one $ l_p $ quasi-norm. At the same time, the new model constructed by it can be transformed into solving two convex sub-models, thereby obtaining an approximate global optimal solution. The coefficients within this hybrid regularization term are obtained through data learning. We apply this new regularization term to sparse optimization by transforming the original problem into two sub-problems. By deriving the corresponding optimality conditions, we develop a novel iterative shrinkage threshold algorithm, called P2ISTA, to solve the smoothed hybrid $ l_p $-$ l_2 $ regularization model. Then the convergence of the algorithm is analyzed. To validate the performance of the proposed algorithm, we conduct computational experiments for well-known image deblurring problems and compare with the existing methods including Soft, Half, PQA, and PLAFPAp. Our computational results clearly demonstrate that the proposed P2ISTA is superior in terms of sparse solution and computational effectiveness.- 1 January 2025
- Published by American Institute of Mathematical Sciences (AIMS) in Journal of Industrial and Management Optimization
- Vol. 21 (7) , 4712-4729
Smoothed hybrid $ l_p $-$ l_2 $ model for sparse optimizationIn this paper, we propose a new regularization term by using a hybrid function of the smoothed sub-one $ l_p $ quasi-norm and $ l_2 $ norm. The new regularization term avoids the non-differentiability of the sub-one $ l_p $ quasi-norm. At the same time, the new model constructed by it can be transformed into solving two convex sub-models, thereby obtaining an approximate global optimal solution. The coefficients within this hybrid regularization term are obtained through data learning. We apply this new regularization term to sparse optimization by transforming the original problem into two sub-problems. By deriving the corresponding optimality conditions, we develop a novel iterative shrinkage threshold algorithm, called P2ISTA, to solve the smoothed hybrid $ l_p $-$ l_2 $ regularization model. Then the convergence of the algorithm is analyzed. To validate the performance of the proposed algorithm, we conduct computational experiments for well-known image deblurring problems and compare with the existing methods including Soft, Half, PQA, and PLAFPAp. Our computational results clearly demonstrate that the proposed P2ISTA is superior in terms of sparse solution and computational effectiveness.- Open Access
- 1 November 2024
- Published by Science China Press., Co. Ltd. in SCIENTIA SINICA Mathematica
基于Eligible正则项的稀疏优化模型与算法及其应用
在当前机器学习、深度学习及统计学习的前沿研究中, 众多核心科学难题能够借助一类基于正则项的稀疏优化模型进行有效表征. 本文致力于深入探索这类基于正则项的稀疏优化模型、有效算法及实际应用场景, 以期推动该领域发展. 为此, 首先系统回顾已有正则项的理论特性, 并创新地提出Eligible正则项的概念. 其次, 建立了基于可分和不可分Eligible正则项的精确恢复理论. 接着, 构建了Eligible正则稀疏优化模型, 并设计了求解该模型的高效邻近梯度算法和加速邻近梯度算法框架. 最后, 通过相位图分析和稀疏指数追踪的实证分析结果显示, 相比传统稀疏模型, Eligible正则稀疏优化模型在保持高度稀疏性的同时, 显著提高了预测精度和模型稳定性.
- 1 November 2024
- Published by Science China Press., Co. Ltd. in SCIENTIA SINICA Mathematica
Open Access基于Eligible正则项的稀疏优化模型与算法及其应用
在当前机器学习、深度学习及统计学习的前沿研究中, 众多核心科学难题能够借助一类基于正则项的稀疏优化模型进行有效表征. 本文致力于深入探索这类基于正则项的稀疏优化模型、有效算法及实际应用场景, 以期推动该领域发展. 为此, 首先系统回顾已有正则项的理论特性, 并创新地提出Eligible正则项的概念. 其次, 建立了基于可分和不可分Eligible正则项的精确恢复理论. 接着, 构建了Eligible正则稀疏优化模型, 并设计了求解该模型的高效邻近梯度算法和加速邻近梯度算法框架. 最后, 通过相位图分析和稀疏指数追踪的实证分析结果显示, 相比传统稀疏模型, Eligible正则稀疏优化模型在保持高度稀疏性的同时, 显著提高了预测精度和模型稳定性.
- 10 June 2024
- Published by Elsevier in ISA Transactions
- Vol. 152, 277-289
Iterative Chebyshev approximation method for optimal control problems
- 10 June 2024
- Published by Elsevier in ISA Transactions
- Vol. 152, 277-289
Iterative Chebyshev approximation method for optimal control problems
- 23 February 2023
- Published by Springer Nature in International Journal of Machine Learning and Cybernetics
- Vol. 14 (8) , 2671-2683
Two-dimensional k-subspace clustering and its applications on image recognition
- 23 February 2023
- Published by Springer Nature in International Journal of Machine Learning and Cybernetics
- Vol. 14 (8) , 2671-2683
Two-dimensional k-subspace clustering and its applications on image recognition
- 1 January 2023
- Published by American Institute of Mathematical Sciences (AIMS) in Journal of Industrial and Management Optimization
- Vol. 19 (5) , 3750-3783
Control parameterization approach to time-delay optimal control problems: A survey
Control parameterization technique is an effective method to solve optimal control problems. It works by approximating the control function by piece-wise constant (or linear) function. In this way, the optimal control problems are approximated by optimal parameter selection problems, which can be regarded as finite-dimensional optimization problems, where the control heights and switching times of the piece-wise constant function are taken as the decision variables. They can be solved by using gradient-based optimization methods. For this, it requires the gradient formulas of the objective and constraint functions with respect to the decision variables. There are two methods for deriving the required gradient formulas. The aim of this paper is to survey various numerical methods developed based on control parameterization for solving optimal control problems governed by time-delay systems. Particular focus is on those numerical methods for optimizing switching time points. This is because the process of optimizing the switching time points is much more complicated for optimal control problems with time delays. In addition, we shall also review the computational methods developed based on control parameterization technique for solving two optimal control problems involving more complicated time delays.
- 1 January 2023
- Published by American Institute of Mathematical Sciences (AIMS) in Journal of Industrial and Management Optimization
- Vol. 19 (5) , 3750-3783
Control parameterization approach to time-delay optimal control problems: A survey
Control parameterization technique is an effective method to solve optimal control problems. It works by approximating the control function by piece-wise constant (or linear) function. In this way, the optimal control problems are approximated by optimal parameter selection problems, which can be regarded as finite-dimensional optimization problems, where the control heights and switching times of the piece-wise constant function are taken as the decision variables. They can be solved by using gradient-based optimization methods. For this, it requires the gradient formulas of the objective and constraint functions with respect to the decision variables. There are two methods for deriving the required gradient formulas. The aim of this paper is to survey various numerical methods developed based on control parameterization for solving optimal control problems governed by time-delay systems. Particular focus is on those numerical methods for optimizing switching time points. This is because the process of optimizing the switching time points is much more complicated for optimal control problems with time delays. In addition, we shall also review the computational methods developed based on control parameterization technique for solving two optimal control problems involving more complicated time delays.
- 1 January 2023
- Published by American Institute of Mathematical Sciences (AIMS) in Journal of Industrial and Management Optimization
- Vol. 19 (2) , 890-915
A new hybrid $ l_p $-$ l_2 $ model for sparse solutions with applications to image processingFinding sparse solutions to a linear system has many real-world applications. In this paper, we study a new hybrid of the \begin{document}$ l_p $\end{document} quasi-norm (\begin{document}$ 0 <p< 1 $\end{document}) and \begin{document}$ l_2 $\end{document} norm to approximate the \begin{document}$ l_0 $\end{document} norm and propose a new model for sparse optimization. The optimality conditions of the proposed model are carefully analyzed for constructing a partial linear approximation fixed-point algorithm. A convergence proof of the algorithm is provided. Computational experiments on image recovery and deblurring problems clearly confirm the superiority of the proposed model over several state-of-the-art models in terms of the signal-to-noise ratio and computational time.- 1 January 2023
- Published by American Institute of Mathematical Sciences (AIMS) in Journal of Industrial and Management Optimization
- Vol. 19 (2) , 890-915
A new hybrid $ l_p $-$ l_2 $ model for sparse solutions with applications to image processingFinding sparse solutions to a linear system has many real-world applications. In this paper, we study a new hybrid of the \begin{document}$ l_p $\end{document} quasi-norm (\begin{document}$ 0 <p< 1 $\end{document}) and \begin{document}$ l_2 $\end{document} norm to approximate the \begin{document}$ l_0 $\end{document} norm and propose a new model for sparse optimization. The optimality conditions of the proposed model are carefully analyzed for constructing a partial linear approximation fixed-point algorithm. A convergence proof of the algorithm is provided. Computational experiments on image recovery and deblurring problems clearly confirm the superiority of the proposed model over several state-of-the-art models in terms of the signal-to-noise ratio and computational time.- 2 December 2022
- Published by Springer Nature in Journal of Global Optimization
- Vol. 86 (2) , 323-353
A non-convex piecewise quadratic approximation of $\ell _{0}$ regularization: theory and accelerated algorithm- 2 December 2022
- Published by Springer Nature in Journal of Global Optimization
- Vol. 86 (2) , 323-353
A non-convex piecewise quadratic approximation of $\ell _{0}$ regularization: theory and accelerated algorithm- 1 March 2022
- Published by American Institute of Mathematical Sciences (AIMS) in Journal of Industrial and Management Optimization
- Vol. 18 (2) , 1305-1320
ON THE OPTIMAL CONTROL PROBLEMS WITH CHARACTERISTIC TIME CONTROL CONSTRAINTS
In this paper, we consider a class of optimal control problems with control constraints on a set of characteristic time instants. By applying the control parameterization technique, these constraints are imposed on the sub-domains that contain the characteristic time points. The values of the control functions as well as the lengths for their corresponding subdomains become decision variables. Time-scaling transformation is an effective technique to optimize the length of each subdomain in a new time horizon. However, the characteristic time instants in the original time horizon become variable time instants in the new time horizon, and hence the control constraints imposed on these characteristic time points are difficult to be formulated in the new time horizon. We propose a surrogate condition and show that the charac-teristic time control constraints will be satisfied once the surrogate condition holds. Moreover, this surrogate condition is easy to formulate in the new time horizon. The resulting approximate problem can be readily solved by many existing computational methods for solving constrained optimal control prob-lems. Finally, we conclude this paper by solving two examples.
- 1 March 2022
- Published by American Institute of Mathematical Sciences (AIMS) in Journal of Industrial and Management Optimization
- Vol. 18 (2) , 1305-1320
ON THE OPTIMAL CONTROL PROBLEMS WITH CHARACTERISTIC TIME CONTROL CONSTRAINTS
In this paper, we consider a class of optimal control problems with control constraints on a set of characteristic time instants. By applying the control parameterization technique, these constraints are imposed on the sub-domains that contain the characteristic time points. The values of the control functions as well as the lengths for their corresponding subdomains become decision variables. Time-scaling transformation is an effective technique to optimize the length of each subdomain in a new time horizon. However, the characteristic time instants in the original time horizon become variable time instants in the new time horizon, and hence the control constraints imposed on these characteristic time points are difficult to be formulated in the new time horizon. We propose a surrogate condition and show that the charac-teristic time control constraints will be satisfied once the surrogate condition holds. Moreover, this surrogate condition is easy to formulate in the new time horizon. The resulting approximate problem can be readily solved by many existing computational methods for solving constrained optimal control prob-lems. Finally, we conclude this paper by solving two examples.
- 5 November 2021
- Published by Springer Nature in Applied Intelligence
- Vol. 52 (8) , 8793-8809
Two-dimensional Bhattacharyya bound linear discriminant analysis with its applications
The recently proposed L2-norm linear discriminant analysis criterion based on Bhattacharyya error bound estimation (L2BLDA) was an effective improvement over linear discriminant analysis (LDA) and was used to handle vector input samples. When faced with two-dimensional (2D) inputs, such as images, converting two-dimensional data to vectors, regardless of the inherent structure of the image, may result in some loss of useful information. In this paper, we propose a novel two-dimensional Bhattacharyya bound linear discriminant analysis (2DBLDA). 2DBLDA maximizes the matrix-based between-class distance, which is measured by the weighted pairwise distances of class means and minimizes the matrix-based within-class distance. The criterion of 2DBLDA is equivalent to optimizing the upper bound of the Bhattacharyya error. The weighting constant between the between-class and within-class terms is determined by the involved data that make the proposed 2DBLDA adaptive. The construction of 2DBLDA avoids the small sample size (SSS) problem, is robust, and can be solved through a simple standard eigenvalue decomposition problem. The experimental results on image recognition and face image reconstruction demonstrate the effectiveness of 2DBLDA.
- 5 November 2021
- Published by Springer Nature in Applied Intelligence
- Vol. 52 (8) , 8793-8809
Two-dimensional Bhattacharyya bound linear discriminant analysis with its applications
The recently proposed L2-norm linear discriminant analysis criterion based on Bhattacharyya error bound estimation (L2BLDA) was an effective improvement over linear discriminant analysis (LDA) and was used to handle vector input samples. When faced with two-dimensional (2D) inputs, such as images, converting two-dimensional data to vectors, regardless of the inherent structure of the image, may result in some loss of useful information. In this paper, we propose a novel two-dimensional Bhattacharyya bound linear discriminant analysis (2DBLDA). 2DBLDA maximizes the matrix-based between-class distance, which is measured by the weighted pairwise distances of class means and minimizes the matrix-based within-class distance. The criterion of 2DBLDA is equivalent to optimizing the upper bound of the Bhattacharyya error. The weighting constant between the between-class and within-class terms is determined by the involved data that make the proposed 2DBLDA adaptive. The construction of 2DBLDA avoids the small sample size (SSS) problem, is robust, and can be solved through a simple standard eigenvalue decomposition problem. The experimental results on image recognition and face image reconstruction demonstrate the effectiveness of 2DBLDA.
- 28 October 2021
- Published by Springer Nature in Journal of the Operations Research Society of China
- Vol. 11 (4) , 983
Publisher Correction to: Intuitionistic Fuzzy Laplacian Twin Support Vector Machine for Semi-supervised Classification
- 28 October 2021
- Published by Springer Nature in Journal of the Operations Research Society of China
- Vol. 11 (4) , 983
Publisher Correction to: Intuitionistic Fuzzy Laplacian Twin Support Vector Machine for Semi-supervised Classification