Which Inverses of Matrices are Invertible?

  • Thread starter Thread starter craigy
  • Start date Start date
  • Tags Tags
    Linear
AI Thread Summary
The discussion focuses on determining the validity of various matrix equations involving invertible nxn matrices A and B. Participants agree that formulas 1 and 3 hold true, while formula 2 is incorrect. The importance of testing with specific examples, such as diagonal matrices, is emphasized to understand the properties of matrix operations better. The identity matrix is clarified as the basis for the calculations, and confusion around notation is addressed. Overall, the consensus is that only formulas 1 and 3 are universally valid for all invertible matrices.
craigy
Messages
4
Reaction score
0
1. Homework Statement

Determine which of the formulas hold for all invertible nxn matrices A and B

1. (In-A)(In+A) = In - A^2
2. (A+B)(A-B) = A^2 - B^2
3. A^8*B^5 is invertible
4. (A+A^-1)^9 = A^9 + A^-9
5. AB = BA
6. A+A^-1 is invertible

I know 5 is right, and number 2 wrong, but what else?
 
Physics news on Phys.org
A starting hint: a product of regular matrices is a regular matrix.
 
I know that but that doesn't solve the problem
 
For 1, 4, and 6, just go ahead and do the multiplication! What is (I-A)(I+A)?
If you don't want to actually multiply (A+A-1) nine times, try one or two times and see if you can't find a pattern. That's how you solve math problems- you try things, you don't just sit and try to "remember" how to solve it! As for (A+ A-1) try to guess an inverse and then do the calculation to see what happens. Try it with a simple diagonal matrix first. (Another general math method- to do complex problems, try a few simple examples first.)
 
I sub n is an identity matrix. 2, 3 and 5 seems to work when using a diagonal matrix.
 
craigy said:
I sub n is an identity matrix. 2, 3 and 5 seems to work when using a diagonal matrix.

You're not interested in diagonal matrices, you're interested in all regular (i.e. invertible) matrices. Take, for example A=\left(\begin{array}{cc}1 & 2\\0 & -1\end{array}\right), and B=\left(\begin{array}{cc}-1 & 0\\2 & 3\end{array}\right). Obviously AB\neq BA, although both det(A) and det(B) are non-zero.

Edit: after reading this, you should easily see if (2) holds or not for two regular nxn matrices A and B.
 
Last edited:
Thanks radou, after some thought, I figured out that 1 and 3 are the only formulas that hold for all invertible n x n matrices A and B.
 
craigy said:
Thanks radou, after some thought, I figured out that 1 and 3 are the only formulas that hold for all invertible n x n matrices A and B.

Assuming that by " (In-A)(In+A) = In - A^2 " you meant " (In-A)(In+A) = In^2 - A^2 ", where In = A^-1.
 
radou said:
Assuming that by " (In-A)(In+A) = In - A^2 " you meant " (In-A)(In+A) = In^2 - A^2 ", where In = A^-1.

In= A-1? Why would you assume that? Craigy specifically said In is the n by n identity matrix.
(In- A)(In+ A)= In2+ InA- AIn+ A2. Since InA= AIn= A, and In2= In, yes, (In- A)(In+ A)= In- A2.
 
  • #10
HallsofIvy said:
In= A-1? Why would you assume that? Craigy specifically said In is the n by n identity matrix.
(In- A)(In+ A)= In2+ InA- AIn+ A2. Since InA= AIn= A, and In2= In, yes, (In- A)(In+ A)= In- A2.

The notation confused me, it's okay now.
 
Back
Top