Hi again,

thanks for your answers on my last posting. I meanwhile have another 
question which you can hopefully help me with.

Here is the scenario:

It is quite easy to calculate the probability p(X > Y), if I have two 
independent normally distributed variables X and Y.

However, I currently don't know how to efficiently calculate the 
probability that one variable is the maximum, if there are more than two 
variables. To be concrete:

X is normally distributed with mu_X and sigma_X

Y_1...Y_n (n Variables) are normally distributed with mu_Y and sigma_Y

mu_X > mu_Y


all variables (X, Y1, ... Yn) are independent!

How can I efficiently calculate the probability that x is the maximum in 
a realisation (x,y1, .... ,y) ?

Well it cannot be calculated by multiplying

p(X>Y1)*p(X>Y2)* .... *p(X>Yn)

I currently calculate the probability that x is the maximum in a 
realisation by transforming the normal distributions to discrete 
distributions. This way I can numerically calculate the searched 
probability. But I guess there is a much better way to do this.

Thanks!

Stefan

.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to