Let ∑ be a set of n × n complex matrices. For m = 1, 2, …, let ∑m be the set of all products of matrices in ∑ of length m. Denote by ∑′ the multiplicative semigroup generated by ∑. ∑ is said to be asymptotically stable (in the sense of dynamical systems) if there is 0 < α < 1 such that there are bounded neighborhoods U, V ⊂ Cn of the origin for which AV ⊂ αmU for all A ∈ ∑m, m = 1, 2, …. For a bounded set ∑ of n × n complex matrices, it is shown that the following conditions are mutually equivalent:
(i) ∑ is asymptotically stable; (ii) \g9(∑) = lim supm → ∞[supA ∈ ∑m ‖ A ‖]1/m < 1; (iii) ϱ(∑) = lim supm → ∞[supA ∈ ∑m ϱ(A)]1/m < 1, where ϱ(A) stands for the spectral radius of A; and (iv) there exists a positive number α such that ϱ(A) ⩽ α < 1 for all A ∈ ∑′. This fact answers an open question raised by Brayton and Tong. The generalized Gelfand spectral radius formula, that is, ϱ(∑) = \g9(∑), conjectured by Daubechies and Lagarias and proved by Berger and Wang using advanced tools from ring theory and then by Elsner using analytic-geometric tools, follows immediately form the above asymptotic stability theorem.
[1]
Robert K. Brayton,et al.
Stability of dynamical systems: A constructive approach
,
1979
.
[2]
L. Elsner.
The generalized spectral-radius theorem: An analytic-geometric proof
,
1995
.
[3]
G. Rota,et al.
A note on the joint spectral radius
,
1960
.
[4]
J. Lagarias,et al.
The finiteness conjecture for the generalized spectral radius of a set of matrices
,
1995
.
[5]
Robert K. Brayton,et al.
Constructive stability and asymptotic stability of dynamical systems
,
1980
.
[6]
Yang Wang,et al.
Bounded semigroups of matrices
,
1992
.
[7]
I. Daubechies,et al.
Sets of Matrices All Infinite Products of Which Converge
,
1992
.