Find all 2x2 matrices such that A=A^-1

  • Thread starter Thread starter autodidude
  • Start date Start date
  • Tags Tags
    Matrices
AI Thread Summary
The discussion revolves around finding all 2x2 matrices A such that A equals its inverse (A = A^-1). It is established that if A = A^-1, then A^2 must equal the identity matrix I, leading to a set of equations derived from the matrix elements. The solutions include the identity matrix and its negative, along with additional matrices that can be expressed in terms of parameters b and c, where specific conditions on a and d apply. Participants express frustration over the complexity of the problem and seek systematic approaches to solving it, highlighting the utility of understanding matrix inversion and eigenvalues in this context. Overall, the exploration reveals various valid matrices that satisfy the initial condition.
autodidude
Messages
332
Reaction score
0

Homework Statement



Find all 2x2 matrices such that A=A^-^1 (the inverse, just in case the notation is different)

Homework Equations



<br /> <br /> A=<br /> \begin{bmatrix}<br /> a &amp; b<br /> \\ c &amp; d<br /> \end{bmatrix}<br />

The Attempt at a Solution



This is my second attempt at this question. The first time, I took a different approach, using the four equations below. This time, I started off with four other different equations and had a very tough time trying to figure it all out. If I hadn't looked at the solutions, I never would've gotten anywhere. Even with it, I struggled a lot and it took an unreasonable amount of time to work it all out...

----

If A=A^-1, then A^2=I. From that, I derived the following equations:

(i)a^2+bc=1
(ii)ab+bd=0
(iii)ac+cd=0
(iv)d^2+bc=1

From (ii) and (iii), if a≠-d, then b=0 and c=0. Then from (i) and (iv),a^2=1 and d^2=1and since a≠-d, then a=1 and d=1 or a=-1 and d=-1

So from that, I get two 2x2 matrices

\begin{bmatrix}
1 & 0
\\ 0&1
\end{bmatrix}

\begin{bmatrix}
-1 & 0
\\ 0&-1
\end{bmatrix}
[/tex]Then to get the others, I let b=0 and c be any real number (again, I never would've done this if I hadn't seen the answers), then to satisfy (iii), a and d would have to be opposite. I did that for c=0 as well and the result was four more matrices

<br /> \begin{bmatrix}<br /> 1 &amp; 0<br /> \\ k &amp;-1<br /> \end{bmatrix}<br /> <br /> \begin{bmatrix}<br /> -1 &amp; 0<br /> \\ k &amp;1<br /> \end{bmatrix}<br /> <br /> \begin{bmatrix}<br /> 1 &amp; k<br /> \\ 0 &amp; -1<br /> \end{bmatrix}<br /> <br /> \begin{bmatrix}<br /> -1 &amp; k<br /> \\ 0&amp;1<br /> \end{bmatrix}<br />

To get the last answer given in the book:

<br /> <br /> \begin{bmatrix}<br /> a &amp; b<br /> \\ \frac{1-a^2}b&amp;-a<br /> \end{bmatrix}<br />

I just rearranged (i) to get b. But couldn't you write represent c in a similar way but with b in the denominator? Are they just writing it in terms of a and b?

Anyway, though it's solved, it took me super long and I'm still not happy with the method I used. I went down another road where I said b=0 iff a+d≠0 because I used

b=\frac{-b}{ad-bc}

ab+bd=0
b(a+d)=0
b=0

And I also said that if b≠0, then ad-bc-1 and I had all these equations but couldn't get anywhere with them.

Basically, I just want to see how you guys would approach this, what would be a systematic way to solve it? This wasn't even a challenge question and I got destroyed by it.

EDIT: Also, I'm unsure about my reasoning. I was more happy with the way I did it last time.
 
Physics news on Phys.org
autodidude said:

Homework Statement



Find all 2x2 matrices such that A=A^-^1 (the inverse, just in case the notation is different)

Homework Equations


<br /> <br /> A=<br /> \begin{bmatrix}<br /> a &amp; b<br /> \\ c &amp; d<br /> \end{bmatrix}<br />

The Attempt at a Solution



If A=A^{-1}, then A^2=I. From that, I derived the following equations:

(i) a^2+bc=1

(ii) ab+bd=0

(iii) ac+cd=0

(iv) d^2+bc=1
...

EDIT: Also, I'm unsure about my reasoning. I was more happy with the way I did it last time.
Subtract equation (iv) from equation (i).

That gives d=\pm\,a\ .

What do equations (ii) & (iii) tell you ?
 
Well there is a quick way of finding the inverse of a 2x2 matrix

A= <br /> \begin{bmatrix} <br /> a &amp; b <br /> \\ c &amp; d <br /> \end{bmatrix}

A^{-1}=\frac{1}{det(A)}*adj(A)

and the adjugate of a 2x2 matrix is just

<br /> \begin{bmatrix} <br /> d &amp; -b <br /> \\ -c &amp; a <br /> \end{bmatrix}

the determinant is ad-bc

which by equating elements of the matrix and it's inverse you can show ad-bc=-1 and that d=-a

and from here getting the rest of the solution isn't too complicated.
 
autodidude said:

Homework Statement



Find all 2x2 matrices such that A=A^-^1 (the inverse, just in case the notation is different)

Homework Equations



<br /> <br /> A=<br /> \begin{bmatrix}<br /> a &amp; b<br /> \\ c &amp; d<br /> \end{bmatrix}<br />

The Attempt at a Solution



This is my second attempt at this question. The first time, I took a different approach, using the four equations below. This time, I started off with four other different equations and had a very tough time trying to figure it all out. If I hadn't looked at the solutions, I never would've gotten anywhere. Even with it, I struggled a lot and it took an unreasonable amount of time to work it all out...

----

If A=A^-1, then A^2=I. From that, I derived the following equations:

(i)a^2+bc=1
(ii)ab+bd=0
(iii)ac+cd=0
(iv)d^2+bc=1

From (ii) and (iii), if a≠-d, then b=0 and c=0. Then from (i) and (iv),a^2=1 and d^2=1and since a≠-d, then a=1 and d=1 or a=-1 and d=-1

So from that, I get two 2x2 matrices

\begin{bmatrix}
1 & 0
\\ 0&1
\end{bmatrix}

\begin{bmatrix}
-1 & 0
\\ 0&-1
\end{bmatrix}
[/tex]


Then to get the others, I let b=0 and c be any real number (again, I never would've done this if I hadn't seen the answers), then to satisfy (iii), a and d would have to be opposite. I did that for c=0 as well and the result was four more matrices

<br /> \begin{bmatrix}<br /> 1 &amp; 0<br /> \\ k &amp;-1<br /> \end{bmatrix}<br /> <br /> \begin{bmatrix}<br /> -1 &amp; 0<br /> \\ k &amp;1<br /> \end{bmatrix}<br /> <br /> \begin{bmatrix}<br /> 1 &amp; k<br /> \\ 0 &amp; -1<br /> \end{bmatrix}<br /> <br /> \begin{bmatrix}<br /> -1 &amp; k<br /> \\ 0&amp;1<br /> \end{bmatrix}<br />

To get the last answer given in the book:

<br /> <br /> \begin{bmatrix}<br /> a &amp; b<br /> \\ \frac{1-a^2}b&amp;-a<br /> \end{bmatrix}<br />

I just rearranged (i) to get b. But couldn't you write represent c in a similar way but with b in the denominator? Are they just writing it in terms of a and b?

Anyway, though it's solved, it took me super long and I'm still not happy with the method I used. I went down another road where I said b=0 iff a+d≠0 because I used

b=\frac{-b}{ad-bc}

ab+bd=0
b(a+d)=0
b=0

And I also said that if b≠0, then ad-bc-1 and I had all these equations but couldn't get anywhere with them.

Basically, I just want to see how you guys would approach this, what would be a systematic way to solve it? This wasn't even a challenge question and I got destroyed by it.

EDIT: Also, I'm unsure about my reasoning. I was more happy with the way I did it last time.

Something you will find useful over and over again is a simple rule for inverting a 2x2 matrix:
swap the diagonal elements, change sign of the off-diagonal elements, and divide by the determinant. (Others have displayed this for you, but expressing it in words may help.)

RGV
 
Hi autodidude! :smile:

If you're interested in alternative and simpler ways to solve it, here's another (although it does require some advanced knowledge :wink:).Since you have ##A^2-I=0##, according to Cayley–Hamilton, the eigenvalues of A must be roots of ##x^2-1=0##.
So the eigenvalues can only be ##\pm 1##.And according to Jordan, the possible matrices A must then be "similar" to one of:
$$
J = \begin{bmatrix}1&0\\0&1\end{bmatrix},\

\begin{bmatrix}-1&0\\0&-1\end{bmatrix},\

\begin{bmatrix}1&0\\0&-1\end{bmatrix},\

\begin{bmatrix}1&1\\0&1\end{bmatrix},\

\begin{bmatrix}-1&1\\0&-1\end{bmatrix}
$$
However not all of these matrices are a solution: the last 2 do not "fit", which you can verify by calculating ##A^2##, or by finding the inverse as mentioned before (swapping the entries on the main diagonal, negating the other 2 entries, and dividing by the determinant).The term "similar" means that for any invertible matrix B, the matrix product ##A = B J B^{-1}## is a solution.

In the cases that J=I or J=-I, you can see that the corresponding matrices simplify to ##A=B I B^{-1}=I##, respectively ##A=B \cdot -I \cdot B^{-1}=-I##.So the solutions with any invertible matrix B are:
$$
A = I,\

A = -I,\

A = B \begin{bmatrix}1&0\\0&-1\end{bmatrix} B^{-1}
$$
 
Last edited:
I did manage to get ad-bc = -1 the second time I did it. I think the part that wasn't supposed to be complicated turned out to be very complicated for me :)

@I like Serena: I'm very interested in other methods...but yeah, that's beyond where I'm at the moment. Looks much nicer :p
 
I tried to combine those 2 formulas but it didn't work. I tried using another case where there are 2 red balls and 2 blue balls only so when combining the formula I got ##\frac{(4-1)!}{2!2!}=\frac{3}{2}## which does not make sense. Is there any formula to calculate cyclic permutation of identical objects or I have to do it by listing all the possibilities? Thanks
Back
Top