Sequential regularizability in the sense of Maslov for improperly posed problems (Q1905296)

The author considers the problem of approximate solution of the equation (1) \(Ax= y\), where \(X\) and \(Y\) are normed spaces and \(A: X\to Y\) is an injective continuous operator with a discontinuous inverse. For solving (1) frequently Galerkin's method or Tikhonov's regularization method is used. Galerkin's method has a remarkable property: If the formally written Galerkin approximations converge, then the limit is necessarily an exact solution. Generally, regularization methods do not have this property. V. P. Maslov first gave a regularizing algorithm in the sense of Tikhonov (or for short TRA) with the property: If the regularized solution converges it follows that it converges precisely to a solution of the equation. The author calls TRA with this property Maslov regularizing algorithm (MRA) and a special kind of it he calls a sequential MRA, which even more than MRA corresponds to the practice of approximate solution of applied problems by the method of regularization. He investigates various properties of the sequential MRA and proves several propositions on them. For instance, a sequential MRA remains a sequential MRA under certain `small' transformations.