Markov-chain Monte-Carlo Sampling for Optimal Fidelity Determination in Dynamic Decision-Making