set condition in R

Exactly! Thanks, Duncan.

Let me re-phrase me question like this:

1) X_i values are independent Gammas, with the shape 0.067 and scale 0.008
2) Min(X)=1 and Max(X)=85
3) SUM(X)=2000
4) Do I also have to define the number of draws? if yes, it could be 250.

Based on these restrictions, I want to generate random draw. I'm wondering
how I can do this in R. Thanks.

Garry



On Tue, Nov 10, 2009 at 11:17 AM, Duncan Murdoch <murdoch at stats.uwo.ca

wrote:

On 11/10/2009 1:25 PM, Hongwei Dong wrote:

Hi, Dear R users,

I'm wondering if I can do Monte Carlo Simulation in R. My problem is

like

this: I know variable X follows Gamma distribution with shape parameter
0.067 and scale parameter 0.008. The sum of the X is 2000. I need R 
help
me
to simulate a vector of X that satisfies both the probability

distribution

and the sum. Anyone has a clue to this? Much appreciated.

Your requirements are slightly contradictory or incomplete.  Here's one

way

to fully specify the problem:

The X_i values are independent Gammas, with the given shape and scale.

You

want to simulate from the joint distribution conditional on the event

sum(X)

== 2000.

Is that your problem?  I don't know how to do the simulation, but maybe
someone else does.

Duncan Murdoch

       [[alternative HTML version deleted]]