{"id":"arxiv/2609.39736","url":"https://arxiv.org/abs/2609.39736","tags":["math.OC"],"score":0,"title":"Improved KKT Complexity for First-Order Bilevel Optimization under Weak Lower-Level Convexity","domain":"math.OC","source":"arxiv","authors":["Jan Harold Alcantara","Masahiro Inoue","Akiko Takeda"],"reviews":0,"subject":"math.OC","summary":"We study deterministic first-order bilevel optimization under weak lower-level convexity, allowing nonconvex lower-level objectives and without assuming strong convexity, the Polyak-Łojasiewicz condition, or an error-bound property. We consider a $δ$-relaxed Moreau-gap constraint, with $δ>0$, for the lower-level stationarity condition and propose an inexact variable-smoothing penalty method (IVSP) for computing its approximate Karush-Kuhn-Tucker (KKT) points. For any fixed relaxation level $δ$, under a standard extended no-nonzero-abnormal-multiplier constraint qualification (ENNAMCQ), we…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-09-30","reactions":{}}