Automated analysis of variance in quantitative simulation be means of Multiple Replications in Parallel

Theoretische Informationstechnik Diplomarbeiten (Mai 2008 - April 2009) it

Betreuer

Rudolf Mathar,

Abstract

Traditional analysis of the output of discrete event simulation concentrates on estimating the mean value of the observations and giving a confidence interval of that estimate, largely because of ease of implementation and analysis. This talk presents new methods for the estimation of the variance of the observations, and their implementation in Akaroa2, an automated controller of quantitative discrete event simulation designed at the University of Canterbury in Christchurch, New Zealand.