Seminar by John Singler

Speaker

John Singler (Missouri University of Science and Technology)

Title

Data approximation capability of proper orthogonal decomposition

Date

  • February 14, 2023 16:00 CET+0100 (Europe/Rome)

  • February 14, 2023 10:00 EST-0500 (US/Eastern)

  • February 14, 2023 09:00 CST-0600 (US/Central)

  • February 14, 2023 07:00 PST-0800 (US/Pacific)

Abstract

In the numerical analysis of Galerkin reduced order models (G-ROMs) based on proper orthogonal decomposition (POD), the data approximation capability of POD plays a crucial role. More specifically, existing approaches to deriving error bounds between the exact solution of the full order model (FOM) and the G-ROM rely on estimates involving the POD data approximation error, i.e., the error between the exact solution data and POD projections of that solution data. We survey advances in POD data approximation theory, and how these results have impacted the numerical analysis of POD G-ROMs for partial differential equations. We consider discrete and continuous data, and different measures of error. We also outline open questions.

Recording

Watch the recording on our YouTube channel.