How to show $\frac{d}{dt}\det(I+At)\mid_{t=0}=\DeclareMathOperator{\Tr}{Tr}\Tr(A)$? I can't find a easy way to prove that.
Asked
Active
Viewed 310 times
0
-
2Did you mean that we should take the value of derivative at $0$? Otherwise this is not true. – zhoraster Sep 11 '15 at 04:17
-
@zhoraster yes, I mean that – 6666 Sep 11 '15 at 04:19
-
1Unless I am mistaken, this is a special case of http://math.stackexchange.com/questions/202535/derivative-of-a-determinant-of-matrix. – Martin R Sep 11 '15 at 04:22
-
@MartinR, yes, but here it is infinitely easier to prove. – zhoraster Sep 11 '15 at 04:24
-
1See this question – GAVD Sep 11 '15 at 04:35
-
@GAVD Thank you very much! – 6666 Sep 11 '15 at 04:41
2 Answers
2
You just need to look at the coefficient before $t$ in the polynomial obtained by expanding the determinant. Now try to identify where the terms $at$ come from. You should note that all the off-diagonal elements contain $t$, so there should not be too many such terms.
You can start with $2\times 2$ or $3\times 3$ to get the idea if you still stuck.

zhoraster
- 25,481
-1
Hint: Note that the characteristic polynomial of $A$ is
$$\det(A-tI)=t^n-\mbox{Tr}(A)t^{n-1}+O(t^{n-2}) \; .$$

Raskolnikov
- 16,108
-
-
1That's really a terrible abuse of notation, since this is correct for unordered fields. Don't use $O(*)$ that way. – Thomas Andrews Sep 11 '15 at 04:29
-