Operational Scheduling of Behind-the-Meter Storage Systems Based on Multiple Nonstationary Decomposition and Deep Convolutional Neural Network for Price Forecasting

Table 1

Shared parameters of neural network models and related strategies selected.

Parameters	MS-CNN
Depth	34
Hidden neural	None
Kernel size	8
Kernel number	24
Batch size	128
Input length	336
Dropout rate	0.05
Loss function	MSE
Optimizing method	AMSGrad
Start learning rate	0.01
Learning rate decay	0.3
Training stop	Early stopping