Bayesian Decision Problems and Markov Chains