Eigenvectors of tensors -- a primer (Q2321923)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Eigenvectors of tensors -- a primer
scientific article

    Statements

    Eigenvectors of tensors -- a primer (English)
    0 references
    27 August 2019
    0 references
    This paper is a primer on eigenvectors of tensors. It is meant to introduce the theory and provide a review of concepts, including some applications. The objects of interest in this study are one-dimensional subspaces of a vector space that are invariant with respect to a homogeneous map of degree \(m>1\). Section 1 of the paper introduces the topic. Section 2 reviews the definition of tensor. A {tensor} of dimension \(n\) and order \(m\geq 1\) over a field \(\mathbb{K}\) (here taken to be either \(\mathbb{R}\) or \(\mathbb{C}\)) is a multilinear map \[ \hat{Q}:\mathbb{K}^n\times \cdots\times \mathbb{K}^n\rightarrow \mathbb{K}^n, \quad (x^{(1)}, \dots, x^{(m)})\mapsto \hat{Q}(x^{(1)}, \dots, x^{(m)}). \] Associated to a tensor \(\hat{Q}\) is the degree-\(m\) homogeneous polynomial map \[ Q:\mathbb{K}^n\rightarrow \mathbb{K}^n, x\mapsto Q(x):=\hat{Q}(x,\dots, x). \] Writing \(x=(x_1,\dots, x_n)^t\), one has \[ Q(x)=\left(\sum_{(i_1, \dots, i_n)}\alpha_{i_1,\dots,i_n}^jx_1^{i_1}\cdots x_n^{i_n}\right)_{1\leq j\leq n} \] where the summation is over all tuples \((i_1, \dots, i_n)\) of nonnegative integers that sum to \(m\). Conversely, the author shows how one may start with a homogeneous polynomial map \(P\) of degree \(m\) and define a symmetric order \(m\) tensor \(\tilde{P}\), thus showing how one may identify symmetric tensors with homogeneous polynomial maps. Section 2 also reviews the definition of eigenvectors and the critical points of homogeneous polynomials. Section 3 focuses on algebraic methods related to eigenvectors of tensors. Specifically, Bezout's theorem in projective space and the application of this theorem to tensors. Section 4 provides an analytic approach to the Brouwer degree of polynomial maps, improves results from Section 3 in the case when the underlying field is \(\mathbb{R}\), and discusses gradient fields. Section 5 discusses applications to polynomial differential equations. Finally, Section 6 discusses the extrema of homogeneous cubic polynomials, which are relevant in the study of the structure theory of liquid crystals.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    Bezout theorem
    0 references
    Brouwer degree
    0 references
    liquid crystal
    0 references
    0 references
    0 references
    0 references
    0 references