Prove the theorem for the matrix

3 replies · 2K views · last reply
frostshoxx
Messages
6
Reaction score
0

Homework Statement



Prove that every square real matrix X can be written in a unique way as the sum of a symmetric matrix A and a skew-symmetric matrix B.

Homework Equations



X = A + B
A = [tex]\frac{X+X^{T}}{2}[/tex]
B = [tex]\frac{X-X^{T}}{2}[/tex]
X = [tex]\frac{X+X^{T}}{2}[/tex] + [tex]\frac{X-X^{T}}{2}[/tex]


The Attempt at a Solution



So I tried to solve [tex]\frac{X+X^{T}}{2}[/tex] + [tex]\frac{X-X^{T}}{2}[/tex] and it gives out X as a solution. However, how can I know that A is a symmetric and B is a skew-symmetric? Any idea?
 
Take the transpose of A and B. You also need to prove uniqueness which I would do by contradiction.
 
Can this be done symbolically? Also, what do you mean by contradiction? could you give some examples?

Thank you for your time.
 
Yes why not, if you take transpose of A, you will get A again. And B is skew because of the negative sign.

Example of uniqueness. Let e be a number (in reals) such that[tex]a \cdot a^{-1}=e[/tex] and [tex]a\cdot e=a \quad \forall a \in \mathbb{R}[/tex]. e is unique.

Proof:
Fix a in reals and assume e is not unique. You have [tex]a\cdot e=a[/tex] and [tex]a\cdot e'=a[/tex] for [tex]e\neq e'[/tex] (same for inverses). Now you have
[tex]a\cdot e \cdot e'=a \cdot e'=a[/tex]
taking inverses gives the result that [tex]e \cdot e'=e[/tex] and [tex]e \cdot e'=e'[/tex]
thus [tex]e=e'[/tex] which contradicts the assumption, thus e must be unique.

Hope that helps