Title :
The spectral radius of a pair of matrices is hard to compute
Author :
Tsitsiklis, John N. ; Blondel, Vincent D.
Author_Institution :
Lab. for Inf. & Decision Syst., MIT, Cambridge, MA, USA
Abstract :
We analyse the computability and the complexity of various definitions of spectral radii for sets of matrices. We show that the joint and generalised spectral radii of two integer matrices are not approximable in polynomial time, and that, two related quantities-the lower spectral radius and the largest Lyapunov exponent-are not algorithmically approximable. As a corollary of these results we show that: 1) the problem of deciding if all possible products of two given matrices are stable is NP-hard, 2) the problem of deciding if at, least one product of two given matrices is stable is undecidable
Keywords :
Lyapunov matrix equations; computability; computational complexity; eigenvalues and eigenfunctions; matrix algebra; stability; NP-hard problem; complexity; computability; eigenvalue; integer matrix pair; largest Lyapunov exponent; lower spectral radius; spectral radius; Computational complexity; Eigenvalues and eigenfunctions; H infinity control; Laboratories; Mathematics;
Conference_Titel :
Decision and Control, 1996., Proceedings of the 35th IEEE Conference on
Conference_Location :
Kobe
Print_ISBN :
0-7803-3590-2
DOI :
10.1109/CDC.1996.573624