A control theory formulation for random variate generation

The need to simulate complex systems in a Monte Carlo manner necessitates efficient methods for generating random variates. We propose a method for random variate generation. The method is based on a control theory formulation. We use a cascade structure consisting of a neural network “controller” and a density estimator (“plant”). The neural network “controller” acts as a density shaper, and is trained until the density of its output (as measured by the density estimator) is as close as possible to the given density. Once training is complete in the design phase, the generation of random numbers can be performed in a very fast manner