Random-subspace Ellipsoidal Trust-region Methods for Derivative-free Optimization
Random-subspace model-based methods offer a promising approach to solving derivative-free optimization (DFO) problems in high dimensions. However, the usual ratio of actual to predicted reduction assesses model agreement at the trial point without directly measuring subspace quality. An accurate prediction in a poorly aligned subspace may therefore trigger trust-region expansion in all directions. We introduce Random-subspace Ellipsoidal TRust-region Optimization (RETRO), a framework for unconstrained DFO using ellipsoidal trust regions. Its trust-region updates combine global size changes…
Comments and reviews
No comments yet.
API
curl /api/v1/records/arxiv/2610.24866
curl -X POST /api/v1/records/arxiv/2610.24866/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2610.24866/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"body": "…", "review": true}'