This paper presents details of an implementation of a computer simulation for the Target Motion Analysis problem where the error of time-delay measurements derived from passively sensed transients is modelled using a Gaussian-mixture probability density function. In order to improve accuracy of the simulation results, a pseudo-random number generator based on shiftregister operations was implemented and compared to the standard C++ random number generator. Von Neumann's accepting-rejection method was used to generate the required non-uniform distribution function. The model was validated and its performance evaluated.