On Some Mean Matrix Inequalites of Dynamical Interest
Let A ∈ SL(n,ℝ). We show that for all n>2 there exist dimensional strictly positive constants C n such that
where ||A|| denotes the operator norm of A (which equals the largest singular value of A), ρ denotes the spectral radius, and the integral is with respect to the Haar measure on O n , normalized to be a probability measure. The same result (with essentially the same proof) holds for the unitary group U n in place of the orthogonal group. The result does not hold in dimension 2. This answers questions asked in [3, 5, 4]. We also discuss what happens when the integral above is taken with respect to measure other than the Haar measure.
KeywordsNeural Network Statistical Physic Complex System Positive Constant Probability Measure
Unable to display preview. Download preview PDF.
- 1.Avila, A., Bochi, J.: A formula with some applications to the theory of Lyapunov exponents. Israel J. Math. 131, 125–137 (2002)Google Scholar
- 2.Brualdi, R., Mellendorf, S.: Regions in the complex plane containing the eigenvalues of a matrix. Am. Math. Monthly 10, 975–985 (1994)Google Scholar
- 3.Burns, K., Pugh, C., Shub, M., Wilkinson, A.: Recent results about stable ergodicity. Proc. Symposia AMS 69, 327–366 (2001)Google Scholar
- 4.Dedieu, J-P., Shub, M.: On random and mean exponents for unitarily invariant and probability measures on Open image in new window Technical report, IBM, 2001. http://www.research.ibm.com/people/s/shub/Dedieu-Shub.pdf, 2001
- 5.Ledrappier, F., Simo, C., Shub, M., Wilkinson, A.: Random versus deterministic exponents in a rich family of diffeomorphisms. J. Stat. Phys. to appearGoogle Scholar
- 6.Hardy, H.G., Littlewood, J., Polya, G.: Inequalities. 2nd Edition, Cambridge: Cambridge University Press, 1988Google Scholar
- 7.Horn, R., Johnson, C.R.: Matrix Analysis. Corrected reprint, Cambridge: Cambridge University Press, England, 1990Google Scholar
- 8.Kato, T.: Perturbation Theory for Linear Operators. Reprint of the 1980 edition, Berlin Heidelberg-New York: Springer, 1995Google Scholar