We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

math.NA

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Numerical Analysis

Title: $\mathcal{L}_2$-optimal Reduced-order Modeling Using Parameter-separable Forms

Abstract: We provide a unifying framework for $\mathcal{L}_2$-optimal reduced-order modeling for linear time-invariant dynamical systems and stationary parametric problems. Using parameter-separable forms of the reduced-model quantities, we derive the gradients of the $\mathcal{L}_2$ cost function with respect to the reduced matrices, which then allows a non-intrusive, data-driven, gradient-based descent algorithm to construct the optimal approximant using only output samples. By choosing an appropriate measure, the framework covers both continuous (Lebesgue) and discrete cost functions. We show the efficacy of the proposed algorithm via various numerical examples. Furthermore, we analyze under what conditions the data-driven approximant can be obtained via projection.
Comments: 22 pages, 10 figures
Subjects: Numerical Analysis (math.NA); Computational Engineering, Finance, and Science (cs.CE); Systems and Control (eess.SY); Optimization and Control (math.OC)
Cite as: arXiv:2206.02929 [math.NA]
  (or arXiv:2206.02929v2 [math.NA] for this version)

Submission history

From: Petar Mlinarić [view email]
[v1] Mon, 6 Jun 2022 22:14:43 GMT (1260kb,D)
[v2] Fri, 14 Oct 2022 15:01:12 GMT (1223kb,D)

Link back to: arXiv, form interface, contact.