Markov Rewards Processes with Impulse Rewards and Absorbing States