My problem is this: I have several mean vectors (evaluated from some mean-function) that are associated with the same covariance matrix. I now what to generate a multivariate normal sample for each such mean vector along with the covariance matrix (that is constant for all the samples).
My solution is to simply iterate (with a for-loop) over the mean vectors and generate the multivariate samples one by one. This works. However, this could be slow. So I am wondering if there is a way to have a collection of mean vectors (represented as matrix, most conveniently) as input to NumPy's numpy.random.multivariate_normal() function?
Is there any other clever way to program this so that it runs fast? Other suggestions?
