Date: Aug 2, 1996 2:07 AM
Author: H.J. Wang
Subject: joint normal distribution


Does anyone know whether there is a built-in function or package that
allows me to write a joint normal distribution in the case when the
individual distributions are NOT independent?

If the answer to the above question is no, then can you tell me how to
integrate, most likely numerically, a function such as Exp[ 1/2(aX^2 +
bX) + Y + c] by X from d to Infinity, where X and Y are variables and a,
b, c, and d are constants.

For the above question, I know the command "Integrate" simply doesn't
work, and "NIntegrate" can not be applied because Y is not a number
(it's a variable). The Gaussian Quadrature in the package
"MumericalMath`GaussianQuadrature`" does not seem to work neither,
because of the upper limit being Infinity.

Any suggestion will be greatly appreciated!

Hung-Jen Wang
PhD Candidate
Economics Department
The University of Michigan