Sensitivity Analysis In Simulation Optimisation
Talal M Alkhamis
Department
of Statistics & Operation Research,
Faculty
of Science, Kuwait University,
P.O.
Box 5969, Safat 13060, KUWAIT.
Abstract
Simulation is
an essential tool for performace evaluation of many practical engineering
systems where engineers typically want to know how the system will perform for
various parameter settings. Since
large-scale simulation may require great amount of computer time and
storage, appropriate statistical
analysis can become quite costly. In
this paper, we develop an interpolating
technique as an effective tool for estimating system respones to parametric
perturbations in simulation. We propose an interpolation
technique where we can simulate at two parameters 
and 
, and interpolate the result for 
. In fact, we can interpolate results for an infinite number
of values of 
. Using this technique the estimates have lower variance than
those produced using Crude Monte Carlo simulation CMC. We are not limited to
interpolating at two points. The
proposed technique can be easily
extended for more than two
values of 
to produce
interpolation estimates. We present computational results to show the validity
of the proposed interpolation technique for sensitivity analysis in
simulation.
