Adaptive action selection using utility-based reinforcement learning | IEEE Conference Publication | IEEE Xplore