A comparative study of 7 algorithms for model reduction

Compares seven model reduction algorithms by applying them to four different dynamical systems. There are four singular value decomposition (SVD) based methods, and three moment matching based methods. The results illustrate that overall, balanced reduction and approximate balanced reduction are the best when we consider whole frequency range. Moment matching methods always lead to higher error norms than SVD based methods due to their local nature; but they are numerically more efficient. Among them, the rational Krylov algorithm gives the best results