Personal tools

Theme for TIFR Centre For Applicable Mathematics, Bangalore

Navigation

You are here: Home / Hurwitz Stability/Instability Analysis of Real Matrices from their Sign Patterns With Applications

# Hurwitz Stability/Instability Analysis of Real Matrices from their Sign Patterns With Applications

Prof. Rama K. Yedavalli Dept. of Mechanical and Aerospace Engineering The Ohio State University, Columbus,
 Speaker Prof. Rama K. Yedavalli Dept. of Mechanical and Aerospace Engineering The Ohio State University, Columbus, Mar 06, 2018 from 04:00 PM to 05:00 PM LH 006 vCal iCal

Abstract:  In this seminar, we address the issue of discerning the Hurwitz stability/instability of a real matrix directly from its sign pattern. In that connection, we first classify all sign patterns of matrices into three categories, namely i) Qualitative Sign Unstable (QLSU) matrices (i.e. matrices with a sign pattern that it is unstable for any magnitudes in the entries, that is, magnitude independent instability), then ii) Qualitative Sign Stable (QLSS) matrices (i.e. matrices with a sign pattern that it is stable for any magnitudes in the entries, that is, magnitude independent stability) and finally iii) MDSU matrices (which require magnitudes to determine its stability/instability). We then propose a necessary and sufficient condition for a sign matrix to be QLSS. The proposed necessary and sufficient condition is derived solely based on the nature (signs) of interactions and interconnections (i.e. based only on the signs of the entries of the matrix), borrowed from ecological principles. The proposed condition in this paper, serves as a better (being much simpler) alternative to the necessary and sufficient condition for QLSS matrices that is available in the ecology literature which uses a complicated test labeled `the Color test'. In addition, few necessary conditions and few sufficient conditions for QLSU/MDSU matrices are presented. Identifying QLSS/MDSU/QLSU sign structures in a necessary and sufficient way has significant implications in many engineering systems whose dynamics are described by linear state space representation.

Filed under: