Error analysis and optimal parameter evaluation in FFT-based 2D-nilt method

Abstract

The paper deals with the method of numerical inversion of two-dimensional Laplace transforms based on a FFT. Its main advantage lies in high speed of calculation, however, it has to be always connected with a proper technique of the convergence acceleration to achieve the required accuracy. It has been shown that either the epsilon or quotient-difference algorithm is suited for this purpose. In the paper the error analysis, comparison and evaluation of optimal NILT parameters are canied out.