基于系统矩阵实Schur分解的集结法模型降阶

AGGREGATION MODEL REDUCTION BASED ON REAL SCHUR DECOMPOSITION OF SYSTEM MATRIX

通过有序实Ｓｃｈｕｒ分解将系统矩阵变成分块对角阵，得到一种数值稳定的集结法模型降阶，并给出降阶的Ｌ∞－误差界．降价系统保留了原系统的主导极点且为最小实现．

In this paper, by transforming the system matrix into a blocked diagonal matrix,it obtains an aggregation model reduction algorithm with numerical stability, and gives a L∞-error bound for model reduction. The reduced-order system retains the prominent poles and its realization is minimal.