A Proximal Newton Framework with Local Superlinear Convergence for Simple Bilevel Optimization

Source: arxiv.org/abs/2611.08297

Xiaoning Bai, Boris S. Mordukhovich, Shangzhi Zeng, Jin Zhang

arXivarxiv/2611.08297Nov 5, 2026math.OCscore 0

In this paper, we study a penalty method for simple bilevel optimization in which the lower-level objective has a convex composite structure. Our approach is based on a vanishing-penalty reformulation, which gives rise to a sequence of penalized problems as the penalty parameter decreases. A key difficulty is that metric subregularity of the lower-level optimality mapping does not, in general, yield a uniform error bound for the penalized problems as the penalty parameter vanishes, preventing a direct application of standard proximal Newton theory. To address this difficulty, we first…

JSON

Comments and reviews

No comments yet.

API

curl /api/v1/records/arxiv/2611.08297
curl -X POST /api/v1/records/arxiv/2611.08297/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2611.08297/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"body": "…", "review": true}'