- #1

- 201

- 0

## Homework Statement

Suppose X

_{1},X

_{2}.... are iid mean 1 exponential random variables. Use large deviation methodology to give a lower bound for the rate function R(a) for a>1

## Homework Equations

R(a) [tex]\leq[/tex] [tex]\frac{-logP[Sn >n*a]}{n}[/tex]

## The Attempt at a Solution

I know that a sum of exponential random variables is Gamma (n, 1).

I'm having a problem with finding the probability.

the pdf of a gamma dist random variable is:[tex]\frac{x^(^n^-^1^)*e^(^-^x^)}{(n-1)!}[/tex]

I think after transforming it, this is the pdf.

But I'm getting a very hard integral and even my TI-89 can't solve it.

I think I'm doing something wrong.

Last edited: