A column generation approach to exact experimental design

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Ahipasaoglu, Selin, Cipolla, Stefano, Gondzio, Jacek
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911686436323328
author Ahipasaoglu, Selin
Cipolla, Stefano
Gondzio, Jacek
author_facet Ahipasaoglu, Selin
Cipolla, Stefano
Gondzio, Jacek
contents In this work, we address the exact D-optimal experimental design problem by proposing an efficient algorithm that rapidly identifies the support of its continuous relaxation. Our method leverages a column generation framework to solve such a continuous relaxation, where each restricted master problem is tackled using a Primal-Dual Interior-Point-based Semidefinite Programming solver. This enables fast and reliable detection of the design's support. The identified support is subsequently used to construct a feasible exact design that is provably close to optimal. We show that, for large-scale instances in which the number of regression points exceeds by far the number of experiments, our approach achieves superior performance compared to existing branch-and-bound-based algorithms in both computational efficiency and solution quality.
format Preprint
id arxiv_https___arxiv_org_abs_2507_03210
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle A column generation approach to exact experimental design
Ahipasaoglu, Selin
Cipolla, Stefano
Gondzio, Jacek
Optimization and Control
In this work, we address the exact D-optimal experimental design problem by proposing an efficient algorithm that rapidly identifies the support of its continuous relaxation. Our method leverages a column generation framework to solve such a continuous relaxation, where each restricted master problem is tackled using a Primal-Dual Interior-Point-based Semidefinite Programming solver. This enables fast and reliable detection of the design's support. The identified support is subsequently used to construct a feasible exact design that is provably close to optimal. We show that, for large-scale instances in which the number of regression points exceeds by far the number of experiments, our approach achieves superior performance compared to existing branch-and-bound-based algorithms in both computational efficiency and solution quality.
title A column generation approach to exact experimental design
topic Optimization and Control
url https://arxiv.org/abs/2507.03210