Simulations of communication systems via integrated variance reduction techniques

Abstract

This paper deals with efficiency issues of estimating common error performance measures of communication systems via Monte Carlo simulations. Efficient simulation techniques using integrated variance reduction techniques (VRTs) are proposed and studied. The integrated techniques are constructed by combining a number of basic VRTs. The general idea is that by integrating several basic VRTs their respective complementary properties can be exploited simultaneously, leading to additional variance reductions of the estimates and, consequently, to substantial gains in terms of simulation time required to obtain them. Based on the basic VRTs, two examples of the integrated techniques are presented and their performance is studied. Simulation results are provided to demonstrate the general approach of the integrated VRTs and to verify the concept. The results indicate that the simulation schemes based on the integrated VRTs can, indeed, improve the efficiency of the simulations. We argue that the great benefit of the integrated techniques is their potential of combining advantages and complementary features of the constituent basic VRTs. Thus, the integrated VRTs can achieve high efficiency gains even though the respective constituent basic VRTs do not operate under their optimal conditions, which is usually the case for a simulation scheme using a single VRT.