Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
NCTM or The Math Forum.



Commuting matrices as polynomials?
Posted:
Feb 18, 2008 2:44 PM


Suppose A and B are commuting nxn matrices. It does not follow that one of these can be expressed as a polynomial in the other. It also does not follow that A and B can both be expressed as polynomials in terms of a third matrix C. In a book by Marcus and Minc (Survey of Matrix Theory and Matrix Inequalities) they give the matrices A=[ [0 0 0], [2 0 0], [0 0 0]] and B=[[0 0 0],[0 0 3], [0 0 4]] as a example (these are given row by row) of the second statement. That is, AB=0 and BA=0 so they commute and Marcus & Minc say that there is no C such that A=p(C) and B=q(C) for polynomials p and q. A bit of hand computation led me to C=[[0 0 0], [1 0 3], [0 0 4]] giving B =1/4 C^2 and A=2C1/2 C^2 Unless I'm missing something it seems that the example given is incorrect. (1) What is the example supposed to be? (2) Can anyone give me a reference to a source that discusses statement 2 (i.e. what is required for A and B to be expressible as polynomials in a third matrix C). I have quite a few linear algebra books but none discuss this)
Thanks jw



