LMRL: a multi-agent reinforcement learning model and algorithm | IEEE Conference Publication | IEEE Xplore