How to Use Reinforcement Learning to Facilitate Future Electricity Market Design? Part 1: A Paradigmatic Theory