Learning-to-solve unit commitment based on few-shot physics-guided spatial-temporal graph convolution network
This work addresses the unit commitment problem in power systems, offering a more efficient and feasible solution method, though it appears incremental as it builds on existing graph convolutional and physics-guided learning techniques.
The paper tackles the unit commitment (UC) problem by proposing a few-shot physics-guided spatial-temporal graph convolutional network (FPG-STGCN) to solve it quickly, achieving better feasibility than mainstream learning methods and higher efficiency than traditional solvers in case studies on IEEE benchmarks.
This letter proposes a few-shot physics-guided spatial temporal graph convolutional network (FPG-STGCN) to fast solve unit commitment (UC). Firstly, STGCN is tailored to parameterize UC. Then, few-shot physics-guided learning scheme is proposed. It exploits few typical UC solutions yielded via commercial optimizer to escape from local minimum, and leverages the augmented Lagrangian method for constraint satisfaction. To further enable both feasibility and continuous relaxation for integers in learning process, straight-through estimator for Tanh-Sign composition is proposed to fully differentiate the mixed integer solution space. Case study on the IEEE benchmark justifies that, our method bests mainstream learning ways on UC feasibility, and surpasses traditional solver on efficiency.