FPL-OPF embeds a fast PF solver in NNs for AC-OPF, using only the last iterations for AD to achieve efficient unsupervised training, with a proof that the resulting gradient approximates the true implicit gradient under mild conditions and experiments showing speedups with near-zero violations.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CE 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Unsupervised Learning for AC Optimal Power Flow with Fast Physics-Aware Layer
FPL-OPF embeds a fast PF solver in NNs for AC-OPF, using only the last iterations for AD to achieve efficient unsupervised training, with a proof that the resulting gradient approximates the true implicit gradient under mild conditions and experiments showing speedups with near-zero violations.