On the regularized inversion of the Laplace transform

Abstract
The authors present an algorithm for the regularized inversion of the Laplace transform. They assume that the values of the Laplace transform are given in a finite number of equidistant points pi and that these values are affected by a Gaussian random noise with zero expectation and standard deviation sigma . The algorithm is applied to several examples and, in general, it gives a good approximation of the normal solution; numerical results are presented for values of sigma in the range (10-6, 10-2) and for sigma =0 (no noise).