Calculate $det'(A)$ using a function $f: mathcalM_2 times 2(mathbbR) to mathcalM_2 times 2(mathbbR)$

The name of the pictureThe name of the pictureThe name of the pictureClash Royale CLAN TAG#URR8PPP











up vote
1
down vote

favorite
1













Problem. Let $mathcalM_2 times 2(mathbbR)$ the space of square matrix of order $2$ and consider the fuction $f: mathcalM_2 times 2(mathbbR) to mathcalM_2 times 2(mathbbR)$ defined by:
$$A=beginpmatrix a&b\c&d\ endpmatrix longmapsto f(A) =beginpmatrix d&-b\-c&a\ endpmatrix.$$
Show that
$$det(A + H) = det(A) + operatornametr(f(A)H) + det(H),$$
and conclude that $det$ is differentiable, with $det'(A) = f(A)^t$. Note that if $det(A) neq 0$, then $f(A) = det(A)A^-1$.




Notation. $f(A)^t$ is the transposed.



I proved that $det(A + H) = det(A) + operatornametr(f(A)H) + det(H)$, but I don't know to use it for show that $det'(A) = f(A)^t$. I tried to write $det(A+H) - det(A) = f(A)^t cdot H + o(H)$ using the equality, and to show that $displaystyle lim_H to 0fraco(H)H = 0$, but I couldn't. I appreciate any hints!







share|cite|improve this question



















  • try expanding $H$ into a linear combination of elementary matrices
    – Max
    Jul 28 at 3:37










  • $|det H | le |H_11 H_22| + |H_12H_21| le 2 |H|^2$. (Cf. Hadamard's inequality.)
    – copper.hat
    Jul 28 at 5:47















up vote
1
down vote

favorite
1













Problem. Let $mathcalM_2 times 2(mathbbR)$ the space of square matrix of order $2$ and consider the fuction $f: mathcalM_2 times 2(mathbbR) to mathcalM_2 times 2(mathbbR)$ defined by:
$$A=beginpmatrix a&b\c&d\ endpmatrix longmapsto f(A) =beginpmatrix d&-b\-c&a\ endpmatrix.$$
Show that
$$det(A + H) = det(A) + operatornametr(f(A)H) + det(H),$$
and conclude that $det$ is differentiable, with $det'(A) = f(A)^t$. Note that if $det(A) neq 0$, then $f(A) = det(A)A^-1$.




Notation. $f(A)^t$ is the transposed.



I proved that $det(A + H) = det(A) + operatornametr(f(A)H) + det(H)$, but I don't know to use it for show that $det'(A) = f(A)^t$. I tried to write $det(A+H) - det(A) = f(A)^t cdot H + o(H)$ using the equality, and to show that $displaystyle lim_H to 0fraco(H)H = 0$, but I couldn't. I appreciate any hints!







share|cite|improve this question



















  • try expanding $H$ into a linear combination of elementary matrices
    – Max
    Jul 28 at 3:37










  • $|det H | le |H_11 H_22| + |H_12H_21| le 2 |H|^2$. (Cf. Hadamard's inequality.)
    – copper.hat
    Jul 28 at 5:47













up vote
1
down vote

favorite
1









up vote
1
down vote

favorite
1






1






Problem. Let $mathcalM_2 times 2(mathbbR)$ the space of square matrix of order $2$ and consider the fuction $f: mathcalM_2 times 2(mathbbR) to mathcalM_2 times 2(mathbbR)$ defined by:
$$A=beginpmatrix a&b\c&d\ endpmatrix longmapsto f(A) =beginpmatrix d&-b\-c&a\ endpmatrix.$$
Show that
$$det(A + H) = det(A) + operatornametr(f(A)H) + det(H),$$
and conclude that $det$ is differentiable, with $det'(A) = f(A)^t$. Note that if $det(A) neq 0$, then $f(A) = det(A)A^-1$.




Notation. $f(A)^t$ is the transposed.



I proved that $det(A + H) = det(A) + operatornametr(f(A)H) + det(H)$, but I don't know to use it for show that $det'(A) = f(A)^t$. I tried to write $det(A+H) - det(A) = f(A)^t cdot H + o(H)$ using the equality, and to show that $displaystyle lim_H to 0fraco(H)H = 0$, but I couldn't. I appreciate any hints!







share|cite|improve this question












Problem. Let $mathcalM_2 times 2(mathbbR)$ the space of square matrix of order $2$ and consider the fuction $f: mathcalM_2 times 2(mathbbR) to mathcalM_2 times 2(mathbbR)$ defined by:
$$A=beginpmatrix a&b\c&d\ endpmatrix longmapsto f(A) =beginpmatrix d&-b\-c&a\ endpmatrix.$$
Show that
$$det(A + H) = det(A) + operatornametr(f(A)H) + det(H),$$
and conclude that $det$ is differentiable, with $det'(A) = f(A)^t$. Note that if $det(A) neq 0$, then $f(A) = det(A)A^-1$.




Notation. $f(A)^t$ is the transposed.



I proved that $det(A + H) = det(A) + operatornametr(f(A)H) + det(H)$, but I don't know to use it for show that $det'(A) = f(A)^t$. I tried to write $det(A+H) - det(A) = f(A)^t cdot H + o(H)$ using the equality, and to show that $displaystyle lim_H to 0fraco(H)H = 0$, but I couldn't. I appreciate any hints!









share|cite|improve this question










share|cite|improve this question




share|cite|improve this question









asked Jul 28 at 3:25









Lucas Corrêa

1,087319




1,087319











  • try expanding $H$ into a linear combination of elementary matrices
    – Max
    Jul 28 at 3:37










  • $|det H | le |H_11 H_22| + |H_12H_21| le 2 |H|^2$. (Cf. Hadamard's inequality.)
    – copper.hat
    Jul 28 at 5:47

















  • try expanding $H$ into a linear combination of elementary matrices
    – Max
    Jul 28 at 3:37










  • $|det H | le |H_11 H_22| + |H_12H_21| le 2 |H|^2$. (Cf. Hadamard's inequality.)
    – copper.hat
    Jul 28 at 5:47
















try expanding $H$ into a linear combination of elementary matrices
– Max
Jul 28 at 3:37




try expanding $H$ into a linear combination of elementary matrices
– Max
Jul 28 at 3:37












$|det H | le |H_11 H_22| + |H_12H_21| le 2 |H|^2$. (Cf. Hadamard's inequality.)
– copper.hat
Jul 28 at 5:47





$|det H | le |H_11 H_22| + |H_12H_21| le 2 |H|^2$. (Cf. Hadamard's inequality.)
– copper.hat
Jul 28 at 5:47











1 Answer
1






active

oldest

votes

















up vote
1
down vote



accepted










Not exactly an answer, but may illuminate slightly.



I guess I would note that the functions $h mapsto det(A+h E_ij)$ are all
polynomials in $h$ and hence smooth. Hence the partial derivatives exist and are
continuous and so $det$ is differentiable.



To find a formula, we can start with special cases.



Note that $det(I+ t E_ij) = 1 + t delta_ij$ and so $D det(I) (E_ij) = delta_ij$.



Hence $D det (I)(H) = sum_ij H_ij det(I)(E_ij) = operatornametr H$.



Now suppose $A$ is invertible. Then $det(A+H) = det A det (I+ A^-1 H)$,
and it follows from this that
$D det(A)(H) = (det A )D det(I)(A^-1H) = det A operatornametr (A^-1H)$. Since
$A operatornameadj(A) = (det A )I$, we have
$D det(A)(H) = operatornametr(operatornameadj(A) H)$.



Since both sides of the latter formula are continuous and the invertible
matrices are dense in the space of matrices, this formula is true for all $A$.



To obtain the other representation of the derivative, we use the inner product
$langle A , B rangle = operatornametr(A^TB)$.
Then $D det(A)(H) = langle operatornameadj(A)^T , H rangle$.






share|cite|improve this answer





















  • Thank you! I understood your idea. So, using $langle A, B rangle = operatornametr(A^TB)$ and write $det(A+H) - det(H) = operatornametr(f(A)H) + det(H)$ and take $o(H) = det(H)$, I can also conclude the same thing, right?
    – Lucas Corrêa
    Jul 28 at 15:12






  • 1




    @LucasCorrêa: Well, there are two different approaches. The one in the answer is a general approach that works for all dimensions. The one in the question is a nice formula for the $2 times 2$ case, and you need to show that $det H$ is $o(|H|)$. If you look at my comment to the question you can see that in fact, $det H$ is $O(|H|^2)$ ('big' O). The formula in the question doesn't easily generalise to higher dimensions.
    – copper.hat
    Jul 28 at 15:16







  • 1




    Yes, I just asked you why I wanted to use the back of the exercise. But your response was very good! In fact, you gave me a much better proof than mine to $det'(A)H = det A operatornametr(A^-1H)$.
    – Lucas Corrêa
    Jul 28 at 15:21










Your Answer




StackExchange.ifUsing("editor", function ()
return StackExchange.using("mathjaxEditing", function ()
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix)
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
);
);
, "mathjax-editing");

StackExchange.ready(function()
var channelOptions =
tags: "".split(" "),
id: "69"
;
initTagRenderer("".split(" "), "".split(" "), channelOptions);

StackExchange.using("externalEditor", function()
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled)
StackExchange.using("snippets", function()
createEditor();
);

else
createEditor();

);

function createEditor()
StackExchange.prepareEditor(
heartbeatType: 'answer',
convertImagesToLinks: true,
noModals: false,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
noCode: true, onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
);



);








 

draft saved


draft discarded


















StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2864956%2fcalculate-deta-using-a-function-f-mathcalm-2-times-2-mathbbr%23new-answer', 'question_page');

);

Post as a guest






























1 Answer
1






active

oldest

votes








1 Answer
1






active

oldest

votes









active

oldest

votes






active

oldest

votes








up vote
1
down vote



accepted










Not exactly an answer, but may illuminate slightly.



I guess I would note that the functions $h mapsto det(A+h E_ij)$ are all
polynomials in $h$ and hence smooth. Hence the partial derivatives exist and are
continuous and so $det$ is differentiable.



To find a formula, we can start with special cases.



Note that $det(I+ t E_ij) = 1 + t delta_ij$ and so $D det(I) (E_ij) = delta_ij$.



Hence $D det (I)(H) = sum_ij H_ij det(I)(E_ij) = operatornametr H$.



Now suppose $A$ is invertible. Then $det(A+H) = det A det (I+ A^-1 H)$,
and it follows from this that
$D det(A)(H) = (det A )D det(I)(A^-1H) = det A operatornametr (A^-1H)$. Since
$A operatornameadj(A) = (det A )I$, we have
$D det(A)(H) = operatornametr(operatornameadj(A) H)$.



Since both sides of the latter formula are continuous and the invertible
matrices are dense in the space of matrices, this formula is true for all $A$.



To obtain the other representation of the derivative, we use the inner product
$langle A , B rangle = operatornametr(A^TB)$.
Then $D det(A)(H) = langle operatornameadj(A)^T , H rangle$.






share|cite|improve this answer





















  • Thank you! I understood your idea. So, using $langle A, B rangle = operatornametr(A^TB)$ and write $det(A+H) - det(H) = operatornametr(f(A)H) + det(H)$ and take $o(H) = det(H)$, I can also conclude the same thing, right?
    – Lucas Corrêa
    Jul 28 at 15:12






  • 1




    @LucasCorrêa: Well, there are two different approaches. The one in the answer is a general approach that works for all dimensions. The one in the question is a nice formula for the $2 times 2$ case, and you need to show that $det H$ is $o(|H|)$. If you look at my comment to the question you can see that in fact, $det H$ is $O(|H|^2)$ ('big' O). The formula in the question doesn't easily generalise to higher dimensions.
    – copper.hat
    Jul 28 at 15:16







  • 1




    Yes, I just asked you why I wanted to use the back of the exercise. But your response was very good! In fact, you gave me a much better proof than mine to $det'(A)H = det A operatornametr(A^-1H)$.
    – Lucas Corrêa
    Jul 28 at 15:21














up vote
1
down vote



accepted










Not exactly an answer, but may illuminate slightly.



I guess I would note that the functions $h mapsto det(A+h E_ij)$ are all
polynomials in $h$ and hence smooth. Hence the partial derivatives exist and are
continuous and so $det$ is differentiable.



To find a formula, we can start with special cases.



Note that $det(I+ t E_ij) = 1 + t delta_ij$ and so $D det(I) (E_ij) = delta_ij$.



Hence $D det (I)(H) = sum_ij H_ij det(I)(E_ij) = operatornametr H$.



Now suppose $A$ is invertible. Then $det(A+H) = det A det (I+ A^-1 H)$,
and it follows from this that
$D det(A)(H) = (det A )D det(I)(A^-1H) = det A operatornametr (A^-1H)$. Since
$A operatornameadj(A) = (det A )I$, we have
$D det(A)(H) = operatornametr(operatornameadj(A) H)$.



Since both sides of the latter formula are continuous and the invertible
matrices are dense in the space of matrices, this formula is true for all $A$.



To obtain the other representation of the derivative, we use the inner product
$langle A , B rangle = operatornametr(A^TB)$.
Then $D det(A)(H) = langle operatornameadj(A)^T , H rangle$.






share|cite|improve this answer





















  • Thank you! I understood your idea. So, using $langle A, B rangle = operatornametr(A^TB)$ and write $det(A+H) - det(H) = operatornametr(f(A)H) + det(H)$ and take $o(H) = det(H)$, I can also conclude the same thing, right?
    – Lucas Corrêa
    Jul 28 at 15:12






  • 1




    @LucasCorrêa: Well, there are two different approaches. The one in the answer is a general approach that works for all dimensions. The one in the question is a nice formula for the $2 times 2$ case, and you need to show that $det H$ is $o(|H|)$. If you look at my comment to the question you can see that in fact, $det H$ is $O(|H|^2)$ ('big' O). The formula in the question doesn't easily generalise to higher dimensions.
    – copper.hat
    Jul 28 at 15:16







  • 1




    Yes, I just asked you why I wanted to use the back of the exercise. But your response was very good! In fact, you gave me a much better proof than mine to $det'(A)H = det A operatornametr(A^-1H)$.
    – Lucas Corrêa
    Jul 28 at 15:21












up vote
1
down vote



accepted







up vote
1
down vote



accepted






Not exactly an answer, but may illuminate slightly.



I guess I would note that the functions $h mapsto det(A+h E_ij)$ are all
polynomials in $h$ and hence smooth. Hence the partial derivatives exist and are
continuous and so $det$ is differentiable.



To find a formula, we can start with special cases.



Note that $det(I+ t E_ij) = 1 + t delta_ij$ and so $D det(I) (E_ij) = delta_ij$.



Hence $D det (I)(H) = sum_ij H_ij det(I)(E_ij) = operatornametr H$.



Now suppose $A$ is invertible. Then $det(A+H) = det A det (I+ A^-1 H)$,
and it follows from this that
$D det(A)(H) = (det A )D det(I)(A^-1H) = det A operatornametr (A^-1H)$. Since
$A operatornameadj(A) = (det A )I$, we have
$D det(A)(H) = operatornametr(operatornameadj(A) H)$.



Since both sides of the latter formula are continuous and the invertible
matrices are dense in the space of matrices, this formula is true for all $A$.



To obtain the other representation of the derivative, we use the inner product
$langle A , B rangle = operatornametr(A^TB)$.
Then $D det(A)(H) = langle operatornameadj(A)^T , H rangle$.






share|cite|improve this answer













Not exactly an answer, but may illuminate slightly.



I guess I would note that the functions $h mapsto det(A+h E_ij)$ are all
polynomials in $h$ and hence smooth. Hence the partial derivatives exist and are
continuous and so $det$ is differentiable.



To find a formula, we can start with special cases.



Note that $det(I+ t E_ij) = 1 + t delta_ij$ and so $D det(I) (E_ij) = delta_ij$.



Hence $D det (I)(H) = sum_ij H_ij det(I)(E_ij) = operatornametr H$.



Now suppose $A$ is invertible. Then $det(A+H) = det A det (I+ A^-1 H)$,
and it follows from this that
$D det(A)(H) = (det A )D det(I)(A^-1H) = det A operatornametr (A^-1H)$. Since
$A operatornameadj(A) = (det A )I$, we have
$D det(A)(H) = operatornametr(operatornameadj(A) H)$.



Since both sides of the latter formula are continuous and the invertible
matrices are dense in the space of matrices, this formula is true for all $A$.



To obtain the other representation of the derivative, we use the inner product
$langle A , B rangle = operatornametr(A^TB)$.
Then $D det(A)(H) = langle operatornameadj(A)^T , H rangle$.







share|cite|improve this answer













share|cite|improve this answer



share|cite|improve this answer











answered Jul 28 at 6:48









copper.hat

122k557156




122k557156











  • Thank you! I understood your idea. So, using $langle A, B rangle = operatornametr(A^TB)$ and write $det(A+H) - det(H) = operatornametr(f(A)H) + det(H)$ and take $o(H) = det(H)$, I can also conclude the same thing, right?
    – Lucas Corrêa
    Jul 28 at 15:12






  • 1




    @LucasCorrêa: Well, there are two different approaches. The one in the answer is a general approach that works for all dimensions. The one in the question is a nice formula for the $2 times 2$ case, and you need to show that $det H$ is $o(|H|)$. If you look at my comment to the question you can see that in fact, $det H$ is $O(|H|^2)$ ('big' O). The formula in the question doesn't easily generalise to higher dimensions.
    – copper.hat
    Jul 28 at 15:16







  • 1




    Yes, I just asked you why I wanted to use the back of the exercise. But your response was very good! In fact, you gave me a much better proof than mine to $det'(A)H = det A operatornametr(A^-1H)$.
    – Lucas Corrêa
    Jul 28 at 15:21
















  • Thank you! I understood your idea. So, using $langle A, B rangle = operatornametr(A^TB)$ and write $det(A+H) - det(H) = operatornametr(f(A)H) + det(H)$ and take $o(H) = det(H)$, I can also conclude the same thing, right?
    – Lucas Corrêa
    Jul 28 at 15:12






  • 1




    @LucasCorrêa: Well, there are two different approaches. The one in the answer is a general approach that works for all dimensions. The one in the question is a nice formula for the $2 times 2$ case, and you need to show that $det H$ is $o(|H|)$. If you look at my comment to the question you can see that in fact, $det H$ is $O(|H|^2)$ ('big' O). The formula in the question doesn't easily generalise to higher dimensions.
    – copper.hat
    Jul 28 at 15:16







  • 1




    Yes, I just asked you why I wanted to use the back of the exercise. But your response was very good! In fact, you gave me a much better proof than mine to $det'(A)H = det A operatornametr(A^-1H)$.
    – Lucas Corrêa
    Jul 28 at 15:21















Thank you! I understood your idea. So, using $langle A, B rangle = operatornametr(A^TB)$ and write $det(A+H) - det(H) = operatornametr(f(A)H) + det(H)$ and take $o(H) = det(H)$, I can also conclude the same thing, right?
– Lucas Corrêa
Jul 28 at 15:12




Thank you! I understood your idea. So, using $langle A, B rangle = operatornametr(A^TB)$ and write $det(A+H) - det(H) = operatornametr(f(A)H) + det(H)$ and take $o(H) = det(H)$, I can also conclude the same thing, right?
– Lucas Corrêa
Jul 28 at 15:12




1




1




@LucasCorrêa: Well, there are two different approaches. The one in the answer is a general approach that works for all dimensions. The one in the question is a nice formula for the $2 times 2$ case, and you need to show that $det H$ is $o(|H|)$. If you look at my comment to the question you can see that in fact, $det H$ is $O(|H|^2)$ ('big' O). The formula in the question doesn't easily generalise to higher dimensions.
– copper.hat
Jul 28 at 15:16





@LucasCorrêa: Well, there are two different approaches. The one in the answer is a general approach that works for all dimensions. The one in the question is a nice formula for the $2 times 2$ case, and you need to show that $det H$ is $o(|H|)$. If you look at my comment to the question you can see that in fact, $det H$ is $O(|H|^2)$ ('big' O). The formula in the question doesn't easily generalise to higher dimensions.
– copper.hat
Jul 28 at 15:16





1




1




Yes, I just asked you why I wanted to use the back of the exercise. But your response was very good! In fact, you gave me a much better proof than mine to $det'(A)H = det A operatornametr(A^-1H)$.
– Lucas Corrêa
Jul 28 at 15:21




Yes, I just asked you why I wanted to use the back of the exercise. But your response was very good! In fact, you gave me a much better proof than mine to $det'(A)H = det A operatornametr(A^-1H)$.
– Lucas Corrêa
Jul 28 at 15:21












 

draft saved


draft discarded


























 


draft saved


draft discarded














StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2864956%2fcalculate-deta-using-a-function-f-mathcalm-2-times-2-mathbbr%23new-answer', 'question_page');

);

Post as a guest













































































Comments

Popular posts from this blog

What is the equation of a 3D cone with generalised tilt?

Color the edges and diagonals of a regular polygon

Relationship between determinant of matrix and determinant of adjoint?