A reinforcement learning method to improve the sweeping efficiency for an agent | IEEE Conference Publication | IEEE Xplore