A government Markov model

We have a bicameral, so we want money to states to match money to districts  Our minimization function is we want a citizen to be equally indifferent to state of federal subsidies. 
Count states up to 50, I think, and count districts by 435.    Let the small counter be the number of pariitons you think in groups of states, it is a test variable, you want to see if all the paritions are in equipartition.  P,S,D and M.  M are payments.  Count payment unit at some precision.  Allocate bandwidth by -iLog(i) as in last time.

This is time independent, it gives you the number of payments you need to make to get the best match.  Then you can scale the unitary payment and do it all at once.