Analysis of adaptability of Reinforcement Learning approach

摘要

Reinforcement Learning is a powerful tool which is being used for solving many optimization problems including power system scheduling problems. Even though there are theoretical results which suggest that under specified technical conditions, RL algorithms are adaptive, however, for power system scheduling problems the potential of adaptability is not still explored. In this paper, we explore, through simulation studies, the adaptability of an RL algorithm considering a simple multi stage decision making problem.