Extension of Tikhonov regularization method using linear fractional programming (Q2297132)

Given a real $m\times n$ matrix $A$ with $m\geq n$ and a noisy vector $b$, this paper focuses on the following least-squares problem: $\min_{x\in \mathbb{R}^n}\|Ax-b\|^2$. Here $\|\cdot\|$ is the Euclidean norm. This problem is commonly studied in the framework of the Tikhonov regularization which consists of solving the regularized least-squares problem: $\min_{x\in \mathbb{R}^n}\|Ax-b\|^2+\|L_{\lambda}x\|^2$, where $L_{\lambda}$ is the regularization matrix and $\lambda$ is the regularization parameter. In recent years, various extensions of the Tikhonov regularization have been proposed where authors used the SVD factorization of the matrix $A$ to provide better-suited regularization matrix and the regularization parameter. Inspired by the recent developments, in this paper the authors propose an extended Tikhonov regularization framework which involves $2n$ unknown parameters. The parameters can be obtained by solving a linear fractional programming problem. Under the appropriate selection of these parameters, various known methods can be subsumed as a particular case of the proposed approach. The authors provide thorough numerical experimentation to show the efficacy and feasibility of the developed regularization framework. This well-written paper may serve as a valuable resource for researchers working in linear inverse and ill-posed problems and in related fields.