1

Let $A\in M_3(\mathbb{R})$ be an orthogonal matrix with $det(A)=1$. Then to show $$(Tr(A)-1)^2+\sum_{i<j}(a_{ij}-a_{ji})^2=4$$ where $Tr$ denotes trace.

Suppose the matrix $A=\left(\begin{matrix} a_{11} & a_{12} & a_{13}\\ a_{21} & a_{22} &a_{23}\\ a_{31} & a_{32} & a_{33} \end{matrix}\right)$ in $M_3(\mathbb{R})$ be orthogonal. Then $Tr(AA^T)=\sum_{i,j}a_{ij}^2=3$. If we expand the left hand side of the equation in question, we have $$LHS=4+2(a_{11}a_{22}+a_{11}a_{33}+a_{22}a_{33})-2(a_{12}a_{21}+a_{13}a_{31}+a_{23}a_{32})-2(a_{11}+a_{22}+a_{33})$$ $$=4+2(a_{11}a_{22}-a_{12}a_{21})+2(a_{11}a_{33}-a_{13}a_{31})+2(a_{22}a_{33}-a_{23}a_{32})-2(a_{11}+a_{22}+a_{33})$$ $$=4+minor(a_{33})+minor(a_{22})+minor(a_{11})-2(a_{11}+a_{22}+a_{33})$$

Is there any theorem/result that can be used to say that the remaining part of the last expression excluding 4 is zero? Or is there any way we can manipulate $a_{ij}$ so that it becomes zero? I tried hard but i see no way out. Any help is highly appreciated.

Dastan
  • 328

1 Answers1

0

I think the LHS should be $$LHS=4+2(minor(a_{33})+minor(a_{22})+minor(a_{11}))−2(a_{11}+a_{22}+a_{33})$$ and you can simplify this expression with the help of the characteristic plynomial of $A$. You can write $$Det(tI-a)=t^3-c_2t^2+c_1t-c_0$$ where $I$ is the identity matrix and $c_i$ is the sum of all principal minor of $A$ of size $3-i$. In particular, $c_0=Det(A)=1$, $c_1=minor(a_{33})+minor(a_{22})+minor(a_{11})$ using your notations and $c_2=Tr(A)$. Finally, $A$ is a matrix of $M_3(\mathbb{R})$ so it has at least one real eigenvalue, $A$ is orthogonal so this eigenvalue is either $1$ or $-1$ and $det(A)=1$ so $A$ has at least an eigenvalue equal to $1$. This means that $Det(I-A)=0$ hence $1-c_2+c_1-c_0=0$ and then $$minor(a_{33})+minor(a_{22})+minor(a_{11})=a_{11}+a_{22}+a_{33}$$ and fially $LHS=4$.

rezoons
  • 421
  • sorry, my bad! I missed the integer 2 – Dastan Nov 15 '18 at 15:00
  • Can you tell me where can I find the proof of the formula of determinant you have used here?Moreover, I din't really get what you meant by saying "A is a matrix of $M_3(\mathbb{R})$ so it has real eigen values," because a real matrix may not have real eigen values. – Dastan Nov 15 '18 at 15:14
  • Is there any way to solve the problem without having resorting to the theory of eigen values? – Dastan Nov 15 '18 at 15:16
  • Anyway, thanks a lot!! – Dastan Nov 15 '18 at 15:18
  • You can find more information on the coefficients of characteristic polynomials with some references here: https://math.stackexchange.com/questions/28650/coefficients-of-characteristic-polynomial-of-a-matrix/144130#144130. In your case, you can always directly expand $Det(tI-A)$ without using this formula.

    Otherwise, i meant that any matrix of $M_3(\mathbb{R})$ have at least one real eigenvalue. That's because the eigenvalues of an $n\times n$ matrix are the roots of the characteristic polynomial with degree $n$ and when $n$ is odd then this polynomial has at least one real eigenvalue.

    – rezoons Nov 15 '18 at 15:40
  • Otherwise, it's possible that you don't need to do all of this to solve your problem. But, i was just answering your question "Is there any theorem/result that can be used to say that the remaining part of the last expression excluding 4 is zero?" and not the problem directly. – rezoons Nov 15 '18 at 15:43
  • I'm absolutely fine with your solution. – Dastan Nov 15 '18 at 16:31