Abstract:Classical model order reduction (MOR) for parametric problems may become computationally inefficient due to large sizes of the required projection bases, especially for problems with slowly decaying Kolmogorov n -widths. Additionally, Hamiltonian structure of dynamical systems may be available and should be preserved during the reduction. In the current presentation, we address these two aspects by proposing a corresponding dictionary-based, online-adaptive MOR approach. The method requires dictionaries for the state-variable, non-linearities, and discrete empirical interpolation (DEIM) points. During the online simulation, local basis extensions/simplifications are performed in an online-efficient way, i.e., the runtime complexity of basis modifications and online simulation of the reduced models do not depend on the full state dimension. Experiments on a linear wave equation and a non-linear Sine-Gordon example demonstrate the efficiency of the approach.

What problem does this paper attempt to address?

The problems that this paper attempts to solve mainly focus on two aspects: 1. **Low computational efficiency of the classical model order reduction method (MOR) when dealing with problems with slowly decaying Kolmogorov n - width**: - Many practical problems cannot be effectively approximated by low - dimensional subspaces, which leads to the need for large projection bases, thus making the model order reduction method computationally inefficient. - The characteristic of this type of problems is that the Kolmogorov n - width of their solution sets decays slowly, so it is difficult for traditional MOR methods to achieve significant acceleration effects. 2. **Model order reduction while maintaining the structure of Hamiltonian systems**: - Hamiltonian systems have a specific structure that can ensure energy conservation and stability. However, classical MOR methods (such as POD) usually destroy this structure, resulting in the reduced - order model may not conserve energy or be unstable. - Therefore, how to maintain the structure of Hamiltonian systems during the model order reduction process is an important research topic. In order to solve the above problems, the paper proposes a dictionary - based online - adaptive structure - preserving model order reduction method (dictionary - based online - adaptive structure - preserving MOR). The main features of this method include: - **Dictionary construction**: In the offline stage, a dictionary containing state variables, nonlinear terms, and discrete empirical interpolation (DEIM) points is pre - constructed. - **Online adaptivity**: During the online simulation process, appropriate bases are selected from the dictionary according to the current parameters and time labels, and local basis expansion or simplification is carried out to maintain computational efficiency. - **Structure preservation**: The specially designed method ensures that the reduced - order model still maintains the structure of the Hamiltonian system, thereby ensuring energy conservation and stability. Specifically, the three main contributions of the paper are: 1. **Introducing a dictionary - based structure - preserving MOR method and proposing the corresponding process**. 2. **Proposing dictionary - based complex singular value decomposition (DB - cSVD)**, which is a new technique for generating symplectic bases. 3. **Efficiently handling nonlinear terms online and introducing the dictionary - based symplectic DEIM (DB - SDEIM) method**. Through these methods, the paper aims to improve the computational efficiency of model order reduction while maintaining the structural characteristics of Hamiltonian systems.

Dictionary-based online-adaptive structure-preserving model order reduction for parametric Hamiltonian systems

Adaptive Sampling for Structure Preserving Model Order Reduction of Port-Hamiltonian Systems

Fully adaptive structure-preserving hyper-reduction of parametric Hamiltonian systems

Structure-Preserving Model Order Reduction for Index Two Port-Hamiltonian Descriptor Systems

Fast Model Order Reduction Via Nonlinear Optimization

Structured Optimization-Based Model Order Reduction for Parametric Systems

Structure-Preserving Model Order Reduction for Index One Port-Hamiltonian Descriptor Systems

A model reduction method for parametric dynamical systems defined on complex geometries

Structure-preserving Model Reduction for Nonlinear Port-Hamiltonian Systems

Model Reduction via Dynamic Mode Decomposition

Symplectic Model Reduction of Hamiltonian Systems on Nonlinear Manifolds

Model Reduction for Large Scale Systems

Model Reduction for Advection Dominated Hyperbolic Problems in an ALE Framework: Offline and Online Phases

A local basis approximation approach for nonlinear parametric model order reduction

Dictionary-based model reduction for state estimation

Randomized low‐rank approximation methods for projection‐based model order reduction of large nonlinear dynamical problems

Non-intrusive data-driven reduced-order modeling for time-dependent parametrized problems

Two-stage model reduction approaches for the efficient and certified solution of parametrized optimal control problems

Parametric Dynamic Mode Decomposition for nonlinear parametric dynamical systems

Structure-Preserving Model Order Reduction for Nonlinear DAE Models of Power Networks