Abstract: Learning-based methods have become mainstream for solving residential energy scheduling problems. In order to improve the learning efficiency of existing methods and increase the utilization ...