Optimal Pursuit Strategies in Discrete-State Probabilistic Systems