English

Polytopes of Stochastic Tensors

Combinatorics 2016-09-14 v1 Functional Analysis

Abstract

Considering n×n×nn\times n\times n stochastic tensors (aijk)(a_{ijk}) (i.e., nonnegative hypermatrices in which every sum over one index ii, jj, or kk, is 1), we study the polytope (Ωn\Omega_{n}) of all these tensors, the convex set (LnL_n) of all tensors in Ωn\Omega_{n} with some positive diagonals, and the polytope (Δn\Delta_n) generated by the permutation tensors. We show that LnL_n is almost the same as Ωn\Omega_{n} except for some boundary points. We also present an upper bound for the number of vertices of Ωn\Omega_{n}.

Keywords

Cite

@article{arxiv.1608.03203,
  title  = {Polytopes of Stochastic Tensors},
  author = {Haixia Chang and Vehbi E. Paksoy and Fuzhen Zhang},
  journal= {arXiv preprint arXiv:1608.03203},
  year   = {2016}
}