Abstract:We develop a unifying framework for interpolatory $\mathcal{L}_2$-optimal reduced-order modeling for a wide classes of problems ranging from stationary models to parametric dynamical systems. We first show that the framework naturally covers the well-known interpolatory necessary conditions for $\mathcal{H}_2$-optimal model order reduction and leads to the interpolatory conditions for $\mathcal{H}_2 \otimes \mathcal{L}_2$-optimal model order reduction of multi-input/multi-output parametric dynamical systems. Moreover, we derive novel interpolatory optimality conditions for rational discrete least-squares minimization and for $\mathcal{L}_2$-optimal model order reduction of a class of parametric stationary models. We show that bitangential Hermite interpolation appears as the main tool for optimality across different domains. The theoretical results are illustrated on two numerical examples.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to develop a unified framework for interpolating L2 - optimal reduced - order modeling, covering a wide range of problems from static models to parametric dynamic systems. Specifically, the objectives of the paper include: 1. **Prove existing interpolation optimality conditions**: For the interpolation optimality conditions in the approximation of linear time - invariant (LTI) systems, prove that these conditions are a special case of its framework and directly derive them therefrom (Section 3.1). 2. **Derive interpolation optimality conditions for parametric LTI systems** (Section 3.2). 3. **Derive interpolation optimality conditions under the discrete least - squares metric** (Section 4). 4. **Derive interpolation optimality conditions for the approximation of parametric static problems** (Section 5). Through these objectives, the authors hope to extend the optimal interpolation theory so that it is applicable to a wider class of problems and provide a unified framework to handle interpolation optimal approximation problems in different fields. ### Specific Problem Description #### 1. Establishment of a Unified Framework The paper first shows that this framework naturally encompasses the known interpolation necessary conditions for H2 - optimal model reduction and leads to the interpolation conditions for H2⊗L2 - optimal model reduction of multi - input/multi - output parametric dynamic systems. In addition, new interpolation optimality conditions for rational discrete least - squares minimization and L2 - optimal model reduction of a class of parametric static models are also derived. #### 2. Bicetare Hermite Interpolation The paper points out that Bicetare Hermite interpolation plays a central role in the optimality in different fields. This indicates that interpolation is a key tool for achieving optimality, whether in dynamic systems or static problems. #### 3. Numerical Examples The theoretical results are illustrated by two numerical examples, which verify the effectiveness and applicability of the proposed framework. ### Main Contributions 1. **Special Case of Existing Conditions**: Proved that the existing interpolation optimality conditions for LTI system approximation are a special case of its framework (Section 3.1). 2. **Conditions for Parametric LTI Systems**: Derived the interpolation optimality conditions for parametric LTI systems (Section 3.2). 3. **Discrete Least - Squares Metric**: Derived the interpolation optimality conditions under the discrete least - squares metric (Section 4). 4. **Parametric Static Problems**: Derived the interpolation optimality conditions for the approximation of parametric static problems (Section 5). Through these contributions, the paper provides a unified framework that can be applied to different types of systems and problems, thereby expanding the application range of the optimal interpolation theory.

A Unifying Framework for Interpolatory $\mathcal{L}_2$-optimal Reduced-order Modeling

Interpolatory $\mathcal{H}_2$-optimality Conditions for Structured Linear Time-invariant Systems

Interpolatory Necessary Optimality Conditions for Reduced-order Modeling of Parametric Linear Time-invariant Systems

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

A New Framework for $\mathcal{H}_2$-Optimal Model Reduction

Interpolatory H-infinity Model Reduction

Interpolatory Model Order Reduction Method for Second Order Systems

Multipoint Volterra Series Interpolation and H2 Optimal Model Reduction of Bilinear Systems

$\mathcal{H}_2$ optimal model reduction of linear systems with multiple quadratic outputs

A Subspace Framework for ${\mathcal L}_\infty$ Model Reduction

Time-limited H2-optimal Model Order Reduction of Linear Systems with Quadratic Outputs

$\mathcal{H}_2$-optimal Model Reduction of Linear Quadratic Output Systems in Finite Frequency Range

H2 optimal rational approximation on general domains

Adaptive choice of near-optimal expansion points for interpolation-based structure-preserving model reduction

Passivity-preserving Model Reduction by Analytic Interpolation

Riemannian Optimization Approach to Structure-Preserving Model Order Reduction of Integral-Differential Systems on the Product of Two Stiefel Manifolds

Optimal H2 moment matching-based model reduction for linear systems by (non)convex optimization

$$\Mathscr{h}_{2}$$ Model Order Reduction for Bilinear Systems Based on the Cross Gramian.

H ∞ Model Reduction of 2-D Singular Roesser Models

Structure-preserving tangential interpolation for model reduction of port-Hamiltonian Systems