Lenient learning in independent-learner stochastic cooperative games

We introduce the Lenient Multiagent Reinforcement Learning 2 (LMRL2) algorithm for independent-learner stochastic cooperative games. LMRL2 is designed to overcome a pathology called relative overge...