Stochastic MDP

The dataset used in this paper is a stochastic MDP with |S| = 4 and |A| = 4. One of the states is set to the terminal state, and one of the rest is set to the starting state. The transition probability and reward functions are randomly generated.