Using Control Variables to Improve the Efficiency of Percentile Estimation in Stochastic Simulation

Control variables may be used in the estimation of percentiles by stochastic stimulation. Hitherto control variables have only been applied to estimating the mean of a response variable; however, in many uses of simulation, information about the tails of the distribution of a response variable are of more interest. The efficacy of the method is demonstrated by application to an example of a simulation study in capital budgeting where exposure to risk is important.