Multiplicities of tensor eigenvalues (Q305077): Difference between revisions

Eigenvalues and eigenvectors of tensors are defined in analogy with those for matrices, as are the algebraic and geometric multiplicities of eigenvalues. But the set of eigenvectors of an eigenvalue of a tensor is not in general a linear subspace, nor is there a simple relationship between the two multiplicities as for matrices. This paper is devoted to the study of this topic. The authors show that the algebraic multiplicity is not invariant under the action of the orthogonal linear group on the tensor, whereas the geometric one is. For generic tensors, however, they show that both multiplicities have the value \(1\). They also propose a general inequality which generalizes the well-known inequality for matrices and holds in a number of interesting special cases.