← ALB News · curated headline

Two-Sided Projection-Based Model Reduction of Parametric Generalized Algebraic Riccati Equations via Grassmann-Manifold Optimization

arXiv Mathematics · WORLD · Mathematics · 11/17/2026, 12:00:00 AM

arXiv:2611.18739v1 Announce Type: new Abstract: Parametric generalized algebraic Riccati equations arise in many-query problems in optimal control, state estimation, and filtering, where repeated full-order solutions may become computationally prohibitive for large-scale systems. In this paper, within a projection-based model reduction framework, we propose a two-sided projection approach that constructs reduced bases by minimizing the aggregate relative two-sided projection error of Riccati solution snapshots. Under suitable dissipativity assumptions, the resulting Galerkin-projected PGARE admits a unique symmetric positive semidefinite stabilizing solution. For error assessment, we employ an existing residual-based a posteriori error estimator for the full-order Riccati residual. A projection-based error decomposition developed in this work provides the theoretical motivation for the proposed basis construction. To accommodate clustered solution manifolds, we extend the framework to localized model reduction by combining parameter-space proximity and solution-space correlation for clustering, with classifier-based online basis selection. Numerical experiments on parameterized heat-conduction problems demonstrate the effectiveness of the proposed global and localized approaches in terms of approximation accuracy and computational efficiency.

Continue to arXiv Mathematics →

You are being forwarded in 6s — stay on this page