Construct a joint probability distribution from correlated variables

Hi, I want to construct the joint probability distribution of two discrete dependent distributions. There is a known correlation between the two input variables. Possibly the solution may be found by the use of copulas.

An example of what I mean: marginal probabilities of two dependent variables a and b: P(a=40,80,100)=[0.25 0.25 0.5] P(b=10,40)=[0.75 0.25] Correlation: f(b)=100-2/3.*a

So how to calculate the joint probability, P(a.*b) ?