Quantity vs. Quality in the Model Order Reduction (MOR) of a Linear System

The goal of any Model Order Reduction (MOR) technique is to build a model of order lower
than the one of the real model, so that the computational effort is reduced, and the ability to estimate the
input-output mapping of the original system is preserved in an important region of the input space. Actually,
since only a subset of the input space is of interest, the matching is required only in this subset of the input
space. In this contribution, the consequences on the achieved accuracy of adopting different reduction
technique patterns is discussed mainly with reference to a linear case study.