Reward processes with nonlinear reward functions