Converse and saturation results on Tikhonov regularization together with a posteriori parameter choice (Q1974192): Difference between revisions

Converse and saturation theorems are proved for the Tikhonov regularization method applied to the nonlinear ill-posed problem \(F(x)= y\) where \(F\) is a weakly closed, continuous, Fréchet differentiable nonlinear operator. For example, the author shows that, under suitable conditions, the optimal convergence rate for the nonlinear Tikhonov method is \(O(\delta^{2/3})\) where \(\delta\) is the error level of the data, if and only if the minimum norm solution \(x_0\) of \(F(x)= y\) lies in the range of the operator \(F'(x_0)^* F'(x_0)\). Order of convergence results are also provided for certain a posteriori parameter choice strategies.