Published December 2019 | Version v1
Journal article

Z-eigenvalues based structured tensors: Mz-tensors and strong Mz-tensors

  • 1. Fudan University, School of Mathematical Sciences (China)
  • 2. Yunnan University, School of Mathematics and Statistics (China)
  • 3. Hexi University, School of Mathematics and Statistics (China)
  • 4. Fudan University, School of Mathematical Sciences and Shanghai Key Laboratory of Contemporary Applied Mathematics (China)

Description

The positive (semi-)definiteness of even-order tensors has been widely studied in these years due to its applications in various aspects, such as spectral hypergraph theory, automatic control, polynomial theory, stochastic process, magnetic resonance imaging and so on. It has been shown that M-tensors, B-tensors, H-tensors, Hilbert tensors and stochastic tensors can be positive definite under proper conditions. However, there are still many positive definite tensors that can not be determined by the above criteria. In this paper, we provide a new class of positive definite tensors whose non-diagonal entries can be positive compared to (strong) M-tensors, and we call it strong Mz-tensors, which can arise from discretizing differential equations, since it is based on Z-eigenvalues rather than H-eigenvalues traditionally. Moreover, we show that an even-order (strong) M-tensor must be an (a strong) Mz-tensor, which reflects the inclusion relationship between even-order M-tensors and Mz-tensors. We also introduce (strong) Hz-tensors, as a generalization of (strong) Mz-tensors, and its positive semi-definiteness (positive definiteness) has been studied. Finally, some conditions are given for a tensor to be an (a strong) Mz-tensor and we use it to study the stability of a high-order nonlinear system.

Additional details

Identifiers

Publishing Information

Journal Title
Computational and Applied Mathematics (Online)
Journal Volume
38
Journal Issue
4
Journal Page Range
p. 1-25
ISSN
1807-0302

INIS

Country of Publication
Brazil
Country of Input or Organization
International Atomic Energy Agency (IAEA)
INIS RN
51081833
Subject category
S97: MATHEMATICAL METHODS AND COMPUTING;
Descriptors DEI
CONTROL; DIFFERENTIAL EQUATIONS; EIGENVALUES; MAGNETIC RESONANCE; NONLINEAR PROBLEMS; POLYNOMIALS; STABILITY; STOCHASTIC PROCESSES; TENSORS
Descriptors DEC
EQUATIONS; FUNCTIONS; RESONANCE

Optional Information

Copyright
Copyright (c) 2019 SBMAC - Sociedade Brasileira de Matematica Aplicada e Computacional