Work

Parametric Interpolation Model Order Reduction on Grassmann Manifolds by Parallelization

Year: 2024

Type: article

Abstract: Based on Riemannian geometry of Grassmann manifolds and discrete Laguerre polynomials, we propose a parametric interpolation parallel MOR method for discrete-time parametric systems. First, a block discrete Fourier transform-based (BDFT) parallel strategy is presented to construct the local basis matrices, which achieves two level acceleration by parallelization. Further, using the retractions and vector transports on Grassmann manifolds, the basis matrix for a new parameter is obtained by interpolating the local bases on the tangent spaces at the different reference points. Finally, a numerical example is used to demonstrate the efficiency of our method.

Source: IEEE Transactions on Circuits & Systems II Express Briefs

DOI: 10.1109/tcsii.2024.3460171

Authors: Kang-Li Xu, Zhen Li, Peter Benner

Institutions: Xi'an Jiaotong University, Xi'an University of Science and Technology, Max Planck Institute for Dynamics of Complex Technical Systems

Language: English

Topic: Model Reduction and Neural Networks

Subfield: Statistical and Nonlinear Physics

Field: Physics and Astronomy

Domain: Physical Sciences

Open Access status: green

Funders: National Natural Science Foundation of China, Fundamental Research Funds for the Central Universities

Awards: G2683882433, G3768268731, G4974434649, G844322515