Convergence analysis of discrete modified Newton scheme for solving ill-posed problems
摘要
Nonlinear inverse problems appear in many applications and can be modelled as an operator equation. In practice, most of these problems are ill-posed, and computing solutions to such problems in an efficient manner is challenging and has been of greatest interest among researchers in the recent past. Generally, iterative methods, particularly Newton-type methods, are among the most effective approaches for solving such problems. However, solving the problem requires the computation of the Fréchet derivative of the nonlinear operator. In traditional iterative methods, this derivative must be calculated at every iteration step, which can be time-consumingTo overcome this, a Frozen Fréchet derivative approach can be employed, thereby significantly reducing computation time. While many approaches are developed within infinite-dimensional Hilbert spaces, practical applications often require solutions in finite-dimensional spaces, where projection methods are commonly used for numerical solutions. In this article, we propose a discrete scheme within the framework of the modified Newton method with Frozen Fréchet derivative to solve nonlinear ill-posed operator equations. We establish convergence under a priori and a posteriori parameter choice rules and derive corresponding error estimates. Apart from this, we also provide numerical examples to illustrate the fact that the proposed scheme is implementable and compare the results with the standard method, showing that our approach yields improved computational performance.