multivariate random variable

Your question is a bit too general to give a useful answer.  One possible answer to your question is:

mrv <- matrix( runif(1000), ncol=10 )

Which generates multivariate random observations, but is unlikely to be what you are really trying to accomplish.  There are many tools for generating multivariate random data including Metropolis-Hastigns, Gibbs sampling, rejection sampling, conditional generation, copulas, and many others, which one will be best (or which combination will be best) depends on what you are actually trying to accomplish.

-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] On Behalf Of Anera Salucci
Sent: Tuesday, November 01, 2011 5:22 AM
To: r-help at r-project.org
Subject: [R] multivariate random variable

Dear All,
?
How can I generate multivariate random variable (not multivariate normal )
?
I am in urgent