Abstract:Under the action of the general linear group with tensor structure, the ranks of matrices $A$ and $B$ forming an $m \times n$ pencil $A + \lambda B$ can change, but in a restricted manner. Specifically, with every pencil one can associate a pair of minimal ranks, which is unique up to a permutation. This notion can be defined for matrix pencils and, more generally, also for matrix polynomials of arbitrary degree. In this paper, we provide a formal definition of the minimal ranks, discuss its properties and the natural hierarchy it induces in a pencil space. Then, we show how the minimal ranks of a pencil can be determined from its Kronecker canonical form. For illustration, we classify the orbits according to their minimal ranks (under the action of the general linear group) in the case of real pencils with $m, n \le 4$. Subsequently, we show that real regular $2k \times 2k$ pencils having only complex-valued eigenvalues, which form an open positive-volume set, do not admit a best approximation (in the norm topology) on the set of real pencils whose minimal ranks are bounded by $2k-1$. Our results can be interpreted from a tensor viewpoint, where the minimal ranks of a degree-$(d-1)$ matrix polynomial characterize the minimal ranks of matrices constituting a block-term decomposition of an $m \times n \times d$ tensor into a sum of matrix-vector tensor products.
What problem does this paper attempt to address?
### What problems does this paper attempt to solve?
This paper aims to explore the minimal ranks of matrix pencils and their applications in tensor decomposition. Specifically, the author mainly solves the following problems:
1. **Definition and Properties**:
- Define the minimal rank of a matrix pencil and study its basic properties.
- Explore how the minimal rank affects the spatial hierarchy of matrix pencils.
2. **Relationship with the Kronecker Canonical Form**:
- Provide a method to determine the minimal rank of a matrix pencil from its Kronecker canonical form.
- Classify the orbits of matrix pencils with different minimal ranks through the Kronecker canonical form.
3. **Applications in Tensor Decomposition**:
- Apply the concept of minimal rank to tensor decomposition, especially the block - term decomposition (BTD).
- Study the relationship between the minimal rank and the tensor rank, revealing their complementary roles in complexity theory.
4. **Approximation Problems**:
- Explore the best approximation problem of real matrix pencils with specific minimal ranks.
- Prove the fact that in some cases there is no best approximation, which is of great significance for applications relying on BTD.
5. **Classification and Examples**:
- For real matrix pencils of \(m\times n\) (\(m, n\leq4\)), completely classify the orbits of the general linear group \(GL_{m,n,2}(\mathbb{R})\) according to their minimal ranks.
- Provide specific examples to illustrate these classification results.
6. **Generalization of Matrix Polynomials**:
- Generalize the concept of minimal rank to matrix polynomials of any finite degree, further expanding its application range.
### Formula Summary
- **Definition of Minimal Rank**:
\[
s=\min_{(t,u)\neq0}\text{rank}(tA + uB)
\]
\[
r=\min_{(t',u')\notin\text{span}\{(t^*, u^*)\}}\text{rank}(t'A + u'B)
\]
where \((t^*, u^*)\) is any pair of parameters that make \(s\) reach the minimum.
- **Minimal Rank under the Kronecker Canonical Form**:
\[
\rho(A,B)=(n - k'_r, n - k'_s)
\]
where \(k'_r\) and \(k'_s\) are the first and second largest components in \((j, k_s, k_r)\) respectively.
- **Minimal Rank of Singular Matrix Pencils**:
\[
\rho(A,B)=(\bar{s}, \bar{s})
\]
where \(\bar{s}=k_1+\cdots + k_p + l_1+\cdots + l_q\).
### Conclusion
This paper, by introducing the concept of the minimal rank of matrix pencils, deeply studies its properties and applications, especially its importance in tensor decomposition and approximation problems. The research results not only enrich matrix theory but also provide theoretical support for practical applications in related fields.