Does anyone know anything (even confirming the existence!) about the multidistribution version of the Fokker Planck equation? For a single distribution \(\rho(x,t)\) I know there's generalisations to go beyond the 2 terms of the FP equation or to have \(x \to \mathbf{x}\) but I'm wondering about having multiple \(\rho_{i}\) which are coupled to one another. Not looking for anything unpleasant (as non-unpleasant as PDEs about distributions can be...), just an overview of basic properties. I can guess what various things about such a concept, if it indeed exists, but I don't trust my gut feelings about probabilities as far as an unfit asthmatic 4 year old can throw an elephant with high cholesterol.
