Multi-Agent Reinforcement Learning And Information Sharing