Multi-parameter Tikhonov regularization (Q356663)

The paper proposes two fixed point algorithms for an error-free parameter choice rule of multiple regularization parameters in Tikhonov regularization suggested in previous papers by the authors. The parameter choice rule is called balancing principle, but it is not related to the Lepskii balancing principle. The convergence of these algorithms is not proved, but fast convergence is shown in numerical experiments. Moreover, the authors derive some properties both of their balancing principle and of the discrepancy priniciple.