How Can We Prove the Inequality of Rank Matrices?

In summary, we have that every column of $AB$ is a linear combination of the columns of $A$, or not? (Wondering). We can do to show the inequality? (Wondering). It follows from the fact that the column space of $AB$ is contained in the column space of $A$, and the row space of $AB$ is contained in the row space of $B$. I'll prove the latter. Given $y$ in the row space of $AB$, there is a row vector $x$ such that $y=xAB$. The vector $z := xA$ is a row vector such that $y=zB$. Thus, $y$ belongs to the row
  • #1
mathmari
Gold Member
MHB
5,049
7
Hey! :eek:

Let $\mathbb{K}$ be a fiels and $A\in \mathbb{K}^{p\times q}$ and $B\in \mathbb{K}^{q\times r}$.
I want to show that $\text{Rank}(AB)\leq \text{Rank}(A)$ and $\text{Rank}(AB)\leq \text{Rank}(B)$.

We have that every column of $AB$ is a linear combination of the columns of $A$, or not? (Wondering)

What can we do to show the inequality? (Wondering)
 
Physics news on Phys.org
  • #2
It follows from the fact that the column space of $AB$ is contained in the column space of $A$, and the row space of $AB$ is contained in the row space of $B$. I'll prove the latter. Given $y$ in the row space of $AB$, there is a row vector $x$ such that $y=xAB$. The vector $z := xA$ is a row vector such that $y=zB$. Thus, $y$ belongs to the row space of $B$.
 
  • #3
Euge said:
Given $y$ in the row space of $AB$, there is a row vector $x$ such that $y=xAB$.

The row space of $AB$ contains the rows of $AB$ that spans the rows, right? (Wondering)
Then $y\in R(AB)$ means that we can write $y$ a linear combination of the vectors of $R(AB)$, right? (Wondering)
We can write this as $y=xAB$ because:
Let $x^T=(x_1, x_2, \ldots , x_{n-1}, x_n)$, then we have that $$xAB=\left (x_1 \cdot (1. \text{ row of }AB))+(x_2 \cdot (2. \text{ row of }AB))+\ldots +(x_{n-1} \cdot ((n-1). \text{ row of }AB))+(x_n \cdot (n. \text{ row of }AB))\right )$$ where some of the $x_i$'s might be $0$.
Is this correct? (Wondering)
 
  • #4
It looks good, although $x^T$ should just be written as $x$, since $x$ is already a row vector.
 
  • #5
Euge said:
It looks good, although $x^T$ should just be written as $x$, since $x$ is already a row vector.

Ah ok!

Let $y\in C(AB)$ then there is a column vector $x$ such that $y=ABx$. The vector $z:Bx$ is a column vector such that $y=Az$. Therefore, $y$ belongs to the column space of $A$.

Is this correct? (Wondering) Having that $C(AB)\subseteq C(A)$ and $R(AB)\subseteq R(B)$, how exactly does it imply that $\text{Rank}(AB)\leq \text{Rank}(A)$ and $\text{Rank}(AB)\leq \text{Rank}(B)$ ? (Wondering)
 
  • #6
Your work is correct. To finish the argument, use the fact that the row rank of a matrix equals its column rank.
 
  • #7
Euge said:
Your work is correct. To finish the argument, use the fact that the row rank of a matrix equals its column rank.

So, we have that $\text{Rank}(AB)=|C(AB)|=|R(AB)|$ and $\text{Rank}(A)=|C(A)|$ and $\text{Rank}(B)=|R(B)|$, right? (Wondering)

Since $C(AB)\subseteq C(A)$, it follows that $|C(AB)|\leq |C(A)|$, or not? (Wondering)

And since $R(AB)\subseteq R(B)$, it follows that $|R(AB)|\leq |R(B)|$.

Therefore, we have that $\text{Rank}(AB)=|C(AB)|\leq |C(A)|=\text{Rank}(A)$ and $\text{Rank}(AB)=|R(AB)|\leq |R(B)|=\text{Rank}(B)$.

Is this correct? (Wondering)
 
  • #8
Looks great!
 
  • #9
Thank you very much! (Smile)
 

FAQ: How Can We Prove the Inequality of Rank Matrices?

1. What is inequality of the rank matrices?

Inequality of the rank matrices refers to the concept that the ranks of two matrices may not be equal, even if the matrices themselves have the same dimensions and elements. This can occur when one matrix has more linearly independent rows or columns than the other.

2. Why is inequality of the rank matrices important?

Inequality of the rank matrices is important because it can help identify relationships between different matrices and determine whether they have the same structure. It can also provide insights into the properties of linear transformations and systems of linear equations.

3. What causes inequality of the rank matrices?

Inequality of the rank matrices can be caused by a variety of factors, such as having different numbers of linearly independent rows or columns, having different sets of pivot elements, or having different numbers of zero rows or columns.

4. Can inequality of the rank matrices be used to solve equations?

Inequality of the rank matrices alone cannot be used to solve equations, but it can provide important information that can aid in solving equations. For example, if two matrices have unequal ranks, it can be concluded that the system of equations represented by the matrices has either no solution or an infinite number of solutions.

5. How can inequality of the rank matrices be determined?

Inequality of the rank matrices can be determined by calculating the ranks of the matrices using row operations or by using the determinant or trace of the matrices. If the ranks are unequal, then inequality of the rank matrices exists.

Similar threads

Replies
7
Views
2K
Replies
9
Views
4K
Replies
9
Views
1K
Replies
10
Views
1K
Replies
4
Views
1K
Replies
16
Views
1K
Replies
3
Views
1K
Back
Top