Lecture Notes for Simulation

23 March 2005 - Output Analysis


IID samples from the random variable X with the underlying n-sample-mean population have the form

X = (Y1 + Y2 + ... + Yn)/n

where Yi is a random variable from the original population (server utilization in the example). The variance of X (that is, vn2) is

Var[X] = Var[(Y1 + Y2 + ... + Yn)/n]

1/n is a constant and can brought outside the variance.

Var[X] = (Var[Y1 + Y2 + ... + Yn])/n2

The variance distributes over addition under the assumption of independent samples.

Var[X] = (Var[Y1] + Var[Y2] + ... + Var[Yn])/n2

By assumption, the variance of the original population is v2, so Var[Yi] = v2 for 1 <= i <= n.

Var[X]=(v2 + v2 + ... + v2)/n2
=(nv2))/n2
=v2/n


This page last modified on 2 March 2005.