Towards possibilistic reinforcement learning algorithms | IEEE Conference Publication | IEEE Xplore