Optimal Policy depending on initial state MDPToolbox Python

Question

I am trying to use MDP Toolbox to implement an algorithm for the "average infinite" reward criteria for a random MDP I have generated through Python\'s MDPToolbox libr