Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Give this matrix A: \begin{pmatrix}-25&2&3&-29\\2&7&7&11\\3&7&7&2\\-29&11&2&11\end{pmatrix}

How can we calculate C matrix when A = AC - CA without extensive computations? Thought of doing these steps:

  1. Let C be {{a,b,c,d},{e,f,g,h},{i,j,k,l},{m,n,o,p}}
  2. Calculate (via wolfram) AC, CA
  3. Subtract AC, CA (cannot do it with wolfram, there's a limit on the characters of input
  4. Then assign the 1st row of the result with the respective row of matrix A
  5. Solve the system of 4 linear equations with 4 variables

Is that correct? How can I find a workaround for step 3?

Thank you for your time!

share|improve this question
    
Why only the look at the first row of the equality A=AC-CA? Wouldn't you need equality in all four rows? Also: wouldn't C be unique only up to addition by matrices B such that AB-BA=0 (ie, the kernel of the adjoint, $\operatorname{ker}\,\operatorname{ad}_A$)? –  anon May 17 '12 at 21:53
    
@anon: "Solve the system of 4 linear equations with 4 variables", that is what I mean, all the 4 rows (I only mentioned the 1st row). –  Chris May 17 '12 at 22:12
add comment

2 Answers

I'm assuming this is homework, so you need to do some work.

First, a good question to ask is how this can be true with scalars (this is can be particularly effective with symmetric matrices)? The answer here is: only if $A$ is zero. This suggests a direction...

Suppose $v$ is a unit eigenvector of $A$ corresponding to the eigenvalue $\lambda$. Consider the value of $v^T A v$: $$ v^T A v = v^T A C v - v^T C A v $$ The left hand side has value $\lambda$, and the right hand side is straightforward to compute.

What does this tell you about $\lambda$?

Since $\lambda$ is an arbitrary eigenvalue, this means all eigenvalues satisfy this condition. What does that tell you about the symmetric matrix $A$?

Given the above, what can you conclude about the solutions of the equation?

share|improve this answer
    
+1 A very nice hint! For some reason I momentarily thought that you need $C$ to be symmetric as well for this to work, but then I reconsidered... –  Jyrki Lahtonen May 17 '12 at 22:25
    
This is a nice hint; however, you seem to assume that $A$ is symmetric. The question doesn't say that explicitly (although the example given is symmetric). –  Fixee May 17 '12 at 22:45
    
I believe it says 'Give [sic.] this matrix'? (Yes, I do assume $A$ is symmetric.) –  copper.hat May 17 '12 at 22:56
    
Why is that "The left hand side has value λ" ? I am new to these concepts and I am a bit confused. But I've thought as a solution to my problem, proving that C exists or not, by using traces only. –  Chris May 18 '12 at 12:53
    
Because $v$ is an eigenvector of $A$ corresponding to $\lambda$ we have $Av = \lambda v$, Since $v$ is a unit vector, $v^T \lambda v = \lambda v^T v = \lambda$. Also, its not clear to me how you would use the trace in this instance. –  copper.hat May 18 '12 at 15:02
show 2 more comments

You can also use vectorization trick: $A=ACI-ICA$ $vec(A)=(A\otimes I)vec(C)-(I\otimes A)vec(C)$ Then C is on one side of the equation so you can solve it.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.