The Minc-type bound and the eigenvalue inclusion sets of the general product of tensors

Chunli Deng,Hongmei Yao,Changjiang Bu
DOI: https://doi.org/10.48550/arXiv.1605.08687
2016-05-27
Abstract:In this paper, we give the Minc-type bound for spectral radius of nonnegative tensors. We also present the bounds for the spectral radius and the eigenvalue inclusion sets of the general product of tensors.
Numerical Analysis
What problem does this paper attempt to address?
The problems that this paper attempts to solve mainly focus on the research of the spectral radius of tensors and the eigenvalue inclusion sets. Specifically: 1. **Minc - type bounds for non - negative tensors**: - The author gives the Minc - type bounds for the spectral radius of non - negative tensors. This bound is a generalization of the classical Minc bound for the matrix spectral radius. - The formula is as follows: \[ \min_{i \in [n]} \frac{r_i(AB)}{(r_i(B))^{m - 1}} \leq \rho(A) \leq \max_{i \in [n]} \frac{r_i(AB)}{(r_i(B))^{m - 1}} \] where \(A\) and \(B\) are \(m\)-order and \(k\)-order tensors respectively. 2. **Generalized Collatz - Wielandt theorem**: - For non - negative weakly irreducible tensors, the author gives the generalized Collatz - Wielandt theorem, further extending the results in the classical matrix theory. - The expression of the theorem is: \[ \max_{B \in R[k,n]_+, r_i(B)\neq 0} \min_{i \in [n]} \frac{r_i(AB)}{(r_i(B))^{m - 1}}=\rho(A)=\min_{B \in R[k,n]_+, r_i(B)\neq 0} \max_{i \in [n]} \frac{r_i(AB)}{(r_i(B))^{m - 1}} \] 3. **Eigenvalue inclusion sets of the generalized product of tensors**: - The paper also studies the spectral radius and eigenvalue inclusion sets of the generalized product of two tensors. - The following formula for the eigenvalue inclusion set is given: \[ \sigma(AB)\subseteq G=\bigcup_{i \in [n]} \left\{z\in \mathbb{C}:|z - c_{ii\ldots i}|\leq r_i(A)(R(B))^{m - 1}-|c_{ii\ldots i}|\right\} \] where \(c_{ii\ldots i}=\sum_{i_2,\ldots,i_m = 1}^n a_{ii_2\ldots i_m}b_{i_2i\ldots i}\ldots b_{imi\ldots i}\). Through these studies, the author aims to provide more accurate upper and lower bounds for the spectral radius of non - negative tensors and explore the eigenvalue distribution laws of the generalized product of tensors. This helps to better understand the properties of high - order tensors and their applications in multilinear algebra.