Signal generator framework-based variation expansion-matching method for polynomial systems.

This paper proposes an interpolatory model reduction method rooted in the signal generator framework for polynomial systems. This innovative approach not only surpasses the constraints on input spaces found in moment-matching methods, such as u (t) ∈ L 2 ([ 0 , + ∞ ]) recommended in H 2 optimal interpolation method, but also eliminates redundant reduced bases through variations of solution in the time domain. Our contributions can be delineated into three key aspects. Firstly, we establish the closure under addition, multiplication, and composition for the transformations from all elementary functions and rational functions to a signal generator-driven system. This validates the applicability of the signal generator framework to general inputs. Secondly, we enhance the numerical stability of two-sided reduced bases for polynomial systems through a variation expansion. Thirdly, we employ tensor decomposition techniques to streamline the complexity of the model reduction process. The simulation results have verified the effectiveness and applicability of the proposed method, which requires smaller computational complexity while maintaining the same level of accuracy. [ABSTRACT FROM AUTHOR]