English

Maximal Exponents of K-Primitive Matrices: The Polyhedral Cone Case

Dynamical Systems 2009-02-27 v1

Abstract

Let KK be a proper (i.e., closed, pointed, full convex) cone in Rn{\Bbb R}^n. An n×nn\times n matrix AA is said to be KK-primitive if there exists a positive integer kk such that Ak(K{0})A^k(K \setminus \{0 \}) \subseteq int KK; the least such kk is referred to as the exponent of AA and is denoted by γ(A)\gamma(A). For a polyhedral cone KK, the maximum value of γ(A)\gamma(A), taken over all KK-primitive matrices AA, is denoted by γ(K)\gamma(K). It is proved that for any positive integers m,n,3nmm,n, 3 \le n \le m, the maximum value of γ(K)\gamma(K), as KK runs through all nn-dimensional polyhedral cones with mm extreme rays, equals (n1)(m1)+1(n-1)(m-1)+1 when mm is even or mm and nn are both odd, and is at least (n1)(m1)(n-1)(m-1) and at most (n1)(m1)+1(n-1)(m-1)+1 when mm is odd and nn is even. For the cases when m=n,m=n+1m = n, m = n+1 or n=3n = 3, the cones KK and the corresponding KK-primitive matrices AA such that γ(K)\gamma(K) and γ(A)\gamma(A) attain the maximum value are identified up to respectively linear isomorphism and cone-equivalence modulo positive scalar multiplication.

Keywords

Cite

@article{arxiv.0902.4589,
  title  = {Maximal Exponents of K-Primitive Matrices: The Polyhedral Cone Case},
  author = {Raphael Loewy and Bit-Shun Tam},
  journal= {arXiv preprint arXiv:0902.4589},
  year   = {2009}
}

Comments

83 pages, 6 figures

R2 v1 2026-06-21T12:15:55.331Z