Maxentropic reconstruction of Fourier and Laplace transforms under nonlinear constraints

摘要

In this paper, the second of a series, we present a maxentropic procedure to recover a continuous function from the knowledge of a few values of a linear integral transform defined by some kernel. The difference from the case treated in the preceding paper lies in the fact that now we assume that the function to be reconstructed takes values in a preassigned interval. We consider as well the case in which the values of the integral transform are known up to some additive noise.