Question: Generate multivariate distribution

Hi,

 

  I think similar question has been asked by several people, but I did not find a suitable thread. My question is, suppose I have a probablity distirubtion function like

  p(x,y) = exp(-alpha (x+y) ) x^2 y^2 / |x-y|  , alpha>0

 x,y goes from - \infty to + \infty. This function is normalizable but unbounded, which makes the rejection algorithm a bit difficult(?).

 

  How to generate samping points from this type of probability distribution function?

 

Thank you very much!

 

Please Wait...