How to Design a Mini Reinforcement Learning Environment-Acting Agent with Intelligent Local Feedback, Adaptive Decision-Making, and Multi-Agent Coordination