On possibility of obtaining linear accuracy evaluation of approximate solutions to inverse problems (Q2364399)

This paper is concerned with solving inverse problems by means of regularization methods. It is well known that generally for ill-posed inverse problems, the error estimate in the noise level of the data is only sublinear. This paper considers the concept of linear a priori estimates of the accuracy of the regularized solution for noisy data. It is shown that linear estimates imply the well-posedness in the sense of Tikhonov. The author also discusses how to obtain the approximation by the method of quasi-solutions and the residual method.