Optimal Strategies in Controlled Markov Chains Under a Stochastic Discount Factor with Applications to Investment Models