Is the matrix A a perfect square?

  • Thread starter Thread starter Maybe_Memorie
  • Start date Start date
  • Tags Tags
    Matrix Square
Join the discussion
Registration is free. Ask a follow-up in this thread, or start your own.
2 replies · 6K views
Maybe_Memorie
Messages
346
Reaction score
0

Homework Statement



Is the following matrix a perfect square? A matrix A is a perfect square if there is a matrix B such that A = B2

____1 1 0 0
A = 0 1 0 0
-----0 0 1 0
-----0 0 0 1

Homework Equations





The Attempt at a Solution



I haven't come across anything like this before, but here's what I'm thinking.

All eigenvalues of A are equal to 1, and the minimal polynomial is m(t) = (t-1)2
and the characteristic polynomial is p(t) = (t-1)4.


Both of these are perfect squares, so does that mean the matrix is a perfect square?
 
Physics news on Phys.org
Maybe_Memorie said:

Homework Statement



Is the following matrix a perfect square? A matrix A is a perfect square if there is a matrix B such that A = B2

____1 1 0 0
A = 0 1 0 0
-----0 0 1 0
-----0 0 0 1

Homework Equations


The Attempt at a Solution



I haven't come across anything like this before, but here's what I'm thinking.

All eigenvalues of A are equal to 1, and the minimal polynomial is m(t) = (t-1)2
and the characteristic polynomial is p(t) = (t-1)4.Both of these are perfect squares, so does that mean the matrix is a perfect square?

Here is a hint: look at the function f(A) = sqrt(e*I + A), where I = identity matrix and e > 0 is a small parameter. You can compute f(A) from the fact that A is a Jordan form with one Jordan block of size 2 and two of size 1, and all eigenvalues are 1. Basically, if f(x) = sum c_n*x^n is an analytic function, then f(A) = c_0*I + sum_{n=1..infinity} c_n * A^n; I = A^0. Since A is so simple, you can get all the A^n easily and do the sum. After you have a nice final formula for f(A) you can see if it has a definite limit as e --> 0. Or, for a quicker way see, eg, Gantmacher, Theory of Matrices, or Google "analytic functions of matrices".

RGV
 
Last edited:
Hi Maybe_Memorie! :smile:

What you're thinking may be true or not.
I do not know (yet).
But without proof you cannot use it.

Here's another method.
If A is a perfect square, it should have one or more square roots.
One method to find a square root is to calculate a power series expansion for a broken power:
[tex](I+X)^{1 \over 2}=I + {1 \over 2} X - {1 \over 8} X^2 + ...[/tex]
Are you already familar with those?