Adaptive dynamic programming for residential energy scheduling with solar energy

摘要

Residential energy scheduling on demand side is a hot research area for saving energy and balancing loads. In this paper, an adaptive dynamic programming method is proposed for residential energy scheduling to reduce cost between two adjacent housing units. Two sets of storage batteries and solar stations make energy scheduling problems quite complicated while using traditional methods. The scheduling algorithm is designed based on action dependent heuristic dynamic programming. In the utility function, the weighting function is given to adjust the remaining capacities of batteries. Furthermore, the temperature becomes an input of neural networks to stay close to reality. Simulation results show the effectiveness of saving cost and balancing loads.