Proving or disproving this matrix V is invertible.

  • Context: MHB 
  • Thread starter Thread starter kaienfr
  • Start date Start date
  • Tags Tags
    Matrix
Click For Summary
SUMMARY

The matrix \( V \) defined by the expression \( V = \begin{bmatrix} (\alpha^1)^{\alpha^1} & \cdots & (\alpha^1)^{\alpha^m}\\ (\alpha^2)^{\alpha^1} & \cdots & (\alpha^2)^{\alpha^m}\\ \vdots & \vdots & \vdots\\ (\alpha^m)^{\alpha^1} & \cdots & (\alpha^m)^{\alpha^m} \end{bmatrix} \) is shown to be invertible under certain conditions. Specifically, when \( n=3 \) and \( d=2 \), the resulting matrix is invertible, as all diagonal elements are strictly positive. However, for cases where \( n=3 \) and \( d=3 \), the matrix is not invertible due to non-zero elements being equal in different positions. The matrix \( V \) is neither triangular nor symmetric in general.

PREREQUISITES
  • Understanding of matrix theory and properties of invertibility.
  • Familiarity with combinatorial concepts, specifically the distribution of balls into bins.
  • Knowledge of matrix operations, including multiplication and determinants.
  • Basic understanding of the implications of diagonal dominance in matrices.
NEXT STEPS
  • Explore the concept of diagonal dominance in matrices and its relation to invertibility.
  • Investigate the properties of non-symmetric matrices and their implications for eigenvalues.
  • Learn about combinatorial matrix theory, particularly in the context of binomial coefficients.
  • Study examples of matrices that are invertible under specific conditions and analyze their structure.
USEFUL FOR

Mathematicians, students studying linear algebra, and researchers in combinatorial optimization will benefit from this discussion, particularly those interested in matrix properties and their applications in combinatorial problems.

kaienfr
Messages
2
Reaction score
0
Hello everyone,
I find an interesting matrix which seems to be always invertible. But I have no idea how to prove it! So I write down here for some ideas. Here is the problem:

Let us take $n\in \mathbb{N}^*$ bins and $d\in \mathbb{N}^*$ balls. Denote the set $B = \{\alpha^1, \ldots, \alpha^m\}$ to be all possible choices for putting $d$ balls into $n$ bins (empty bin is possible), such as
$$\alpha^1 = (d,0,\ldots, 0), ~ \alpha^2 = (0,d,\ldots, 0), \ldots$$
Let us define the matrix $V$ as:
$$V = \begin{bmatrix}
(\alpha^1)^{\alpha^1} & \cdots & (\alpha^1)^{\alpha^m}\\
(\alpha^2)^{\alpha^1} & \cdots & (\alpha^2)^{\alpha^m}\\
\vdots & \vdots & \vdots\\
(\alpha^m)^{\alpha^1} & \cdots & (\alpha^m)^{\alpha^m}
\end{bmatrix}$$
where the notation $(\alpha^i)^{\alpha^j} = \displaystyle\prod_{k=1}^{n}(\alpha^i_k)^{\alpha^j_k}$, under the assumption that $0^0=1$.

Question: "is the matrix $V$ invertible?"

I have tested several examples, and it seems that $V$ is always invertible, but I have no idea how to prove it or find a counterexample.
Here are two facts I understood:
  1. all diagonal elements are strictly positives, so the trace of $V$ is strictly positive.
  2. the matrix $V$ is not symmetric.
Does anyone can help to prove or disprove the invertibility of $V$? Thanks a lot in advance for sharing any idea.
 
Physics news on Phys.org
For a better understanding of the problem, here is an example:

Let $n=3$ and $d=2$, then we have all possible choices for putting $2$ balls into $3$ bins as:
$$B = \{(2,0,0),(0,2,0),(0,0,2),(1,1,0),(1,0,1),(0,1,1)\}.$$
The elements in the first row of the matrix $V$ are computed as:
$$(\alpha^1)^{\alpha^1} = (2,0,0)^{(2,0,0)} = 2^2\times 0^0\times 0^0 = 4,$$
$$(\alpha^1)^{\alpha^2} = (2,0,0)^{(0,2,0)} = 2^0\times 0^2\times 0^0 = 0,$$
and so on.
Thus, we have the matrix $V$:
$$V = \begin{bmatrix}
4&0&0&0&0&0\\
0&4&0&0&0&0\\
0&0&4&0&0&0\\
1&1&0&1&0&0\\
1&0&1&0&1&0\\
0&1&1&0&0&1\\
\end{bmatrix}
$$
which is clearly invertible.

Remarks: Note that, the matrix $V$ is not always triangular. as an example, for n=3, d=3, then the element $$(2,1,0)^{(1,2,0)} = 2^1\times 1^2 \times 0^0 = 2,$$ and $$(1,2,0)^{(2,1,0)} = 1^2\times 2^1 \times 0^0 = 2$$
which are both non-zeros, thus $V$ will be never invertible in this case.

Moreover, when the elements $(\alpha^i)^{\alpha^j}$ and $(\alpha^j)^{\alpha^i}$ (with $i\neq j$) are non-zeros, they may not be equal.
E.g., n=4, d=8, we have
$$(1,1,2,4)^{(2,2,2,2)} = 64$$
but
$$(2,2,2,2)^{(1,1,2,4)} = 256.$$

So, as a conclusion: "the matrix $V$ is neither triangular nor symmetric in general."
 

Similar threads

  • · Replies 52 ·
2
Replies
52
Views
4K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 9 ·
Replies
9
Views
2K
  • · Replies 15 ·
Replies
15
Views
5K
  • · Replies 34 ·
2
Replies
34
Views
3K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 21 ·
Replies
21
Views
1K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 23 ·
Replies
23
Views
2K