Enhanced augmented Lagrangian Hopfield network for unit commitment
An enhanced augmented Lagrangian Hopfield network (EALHN) for unit commitment (UC) is proposed. The EALHN is an augmented Lagrangian Hopfield network (ALHN) enhanced by unit classification to allow decommitment of excess spinning reserve units caused by the minimum up and down time constraints. First, the ALHN is used to solve the UC problem when neglecting the minimum up and down time constraints. Then, a heuristic-based algorithm is applied to satisfy the minimum up and down time constraints and decommit excess spinning reserve units. Finally, the economic dispatch problem is solved using an augmented Lagrangian-relaxation-based continuous Hopfield network (ALRHN). The EALHN is tested on several systems ranging from 10 to 100 units and compared to several methods. The total production costs from the proposed method are smaller those obtained using existing methods, especially for the large systems. Moreover, the computational times of the proposed method are also much faster than these for existing methods and slightly increase with the system size, which is very favourable for large-scale implementation.
