Multi-Agent Reinforcement Learning: A Selective Overview of Theories and Algorithms