Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

What are values of $a,b,c$ such that matrix $A$ and $B$ commutes, i.e. $AB=BA$, where $A$ and $B$ are (3,3) matrix such that $A=\pmatrix{a_0 & b_0 &c_0\\ a_1 & b_1 & c_1\\ a_2 & b_2 & c_2}$

$$B=\pmatrix{1 & 1 &1\\ 1 & \omega & \omega^2\\ 1 & \omega^2 & \omega}$$ where $\omega+\omega^2+1=0$

share|cite|improve this question
If I understand your question correctly, isn't $$A=\pmatrix{0 & 0 &0\\ 0 & 0 &0\\ 0& 0 & 0}$$ a valid answer, because $$AB=\pmatrix{0 & 0 &0\\ 0 & 0 &0\\ 0& 0 & 0}=BA?$$ Or is the question asking you to find every matrix $A$ for which $A$ and $B$ commute? – Zev Chonoles Jun 19 '11 at 5:16
Thanks for LaTeX. Yes, identity or zero matrix satisfies the condition but I want to find general form. – user12290 Jun 19 '11 at 5:37
You can define a 9-component vector v composed of those a, b, c coefficients, then write AB-BA=0 as a linear system Xv=0, and then v is any vector in the kernel space of X. Also, you might need to specify that $ \omega $ is one of $ \exp(2\pi i / 3) $ or $ \exp(4\pi i /3) $ and not the other, as they are both solutions to the quadratic. – anon Jun 19 '11 at 5:42
up vote 8 down vote accepted

Yes, as the other answer notes, B has distinct eigenvalues. Therefore, anything commuting with it must be a polynomial in B. (Degrees no higher than 2 suffice.)

In some detail: let T be a matrix so that $TBT^{-1}$ is diagonal, with distinct diagonal entries $b_1, b_2, b_3$. Since $AB=BA$, $(TAT^{-1})(TBT^{-1})=(TBT^{-1})(TAT^{-1})$. For any matrix $M$, the $ij$ entry of $M(TBT^{-1})$ is $m_{ij}b_j$, while the $ij$ entry of $(TBT^{-1})M$ is $b_im_{ij}$. Thus, a matrix commuting with $TBT^{-1}$ is diagonal. Use Lagrange interpolation to find a quadratic polynomial $P$ so that $P(TBT^{-1})=TAT^{-1}$ is that other polynomial. Happily, $P(TBT^{-1})=T.P(B).T^{-1}$, so $A=P(B)$.

share|cite|improve this answer
It is clear that polynomial in B must be commutative with B. Will you kindly explain why converse is true? – user12290 Jun 20 '11 at 14:47
@user12290: Let the eigenvalues of $B$ be $\lambda_i, i=1,2,3$. Anything that commutes with $B$ must map an eigenspace of $B$ to itself: if $Bx=\lambda_i x$, then $BAx=ABx=\lambda_i Ax$. Here the eigenspaces are 1-dimensional, so $A$ must act on every one of them as a scalar. IOW $A$ is determined by the three scalars $\mu_i,i=1,2,3$. By Lagrange's interpolation formula there exists a quadratic polynomial $p(x)$ such that $p(\lambda_i)=\mu_i$ for all $i=1,2,3$. Thus $A=p(B)$. – Jyrki Lahtonen Jun 21 '11 at 13:05
Thanks for the details. – user12290 Jun 24 '11 at 5:16

It turns out that $B$ has three different eigenvalues, so in a suitable basis it assumes diagonal form $D$. If $\hat A$ is the correspondingly transformed matrix $A$ then $AB=BA$ iff $\hat A D=D\hat A$. Now you have a simpler problem.

As you can see in Paul Garrett's answer you don't even have to compute the eigenvectors of $B$ and get a parametric representation of all $A$'s commuting with $B$ for free.

share|cite|improve this answer
Will you kindly give the details? – user12290 Jun 20 '11 at 11:49

Your Answer


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.