1. ## Matrix Polynomials Question

I'm studying for a PhD preliminary exam, and it seems I have some holes in my linear algebra knowledge. I need help with the following problem:

Let A and B be operators V -> V, where V is a finite-dimensional vector space over the complex numbers. Let p be any polynomial such that p(AB) = 0. Then (a) Prove that that q(x) = xp(x) satisfies q(BA) = 0.

Interestingly, I have shown that this problem is equivalent to proving that the minimal polynomial of AB is equal to the minimal polynomial of BA. Therefore, a direct solution to this problem would be to directly show that q(BA) = 0 whereas, given what I've already shown, it would suffice to simply prove that the minimal polynomials of BA and AB are equal. I am interested to a solution of either kind (ideally both!)

Thanks!

2. ## Re: Matrix Polynomials Question

I think the fact that this problem is about operators over finite-dimensional vector space is irrelevant. This statement can be proved about any associative algebra over a ring.

3. ## Re: Matrix Polynomials Question

suppose p(x) = c0 + c1x + ...+ cmxm, so that

q(x) = c0x + c1x2 +....+ cmxm+1.

then q(BA) = c0BA + c1(BA)2 +...+ cm(BA)m+1.

note that (BA)k = B(AB)k-1A, so we have:

q(BA) = B(c0I + c1(AB) +.... + cm(AB)m)A

= B(p(AB))A = B(0)A = 0.

your assertion that AB and BA must have the same minimal polynomial is false. suppose A =

[1 -1]
[1 -1], and that B =

[0 1]
[0 1]

then AB = 0, while BA = A.

clearly, the minimal polynomial for AB is x, while the minimal polynomial for BA is x2. note this agrees with the result above.