Future Polynomial Regularization of Ill-Posed Volterra Equations (Q4943650)

The authors propose a discrete method for regularizing ill-posed Volterra problems associated to linear integral equations of the first kind. Unlike the classical regularization technique of Tikhonov, this method preserves the Volterra (causal) nature of the problem allowing the regularized solution to be produced sequentially in near real time as data arrived. Convergence of the regularized solution to the true solution is proved at the best possible rate with respect to error in the data. More precisely, the authors use the ``future polynomial regularization method'' that can be applied provided that the operator may be discretized by a lower-triangular matrix of Toeplitz type. The method is based on a local regularization strategy and works for linear equations. Numerical results are given and comparisons with classical methods are discussed.