def reward_function(state, action):
    target_reward = compute_target_reward(state)
    obstacle_penalty = compute_obstacle_penalty(state)
    energy_penalty = compute_energy_penalty(action)
    return target_reward - obstacle_penalty - energy_penalty