3
$\begingroup$

I am trying to find the Smith normal form of matrix over $ \mathbb{R} [ X ]$ of the 4x4 matrix

$$M =\begin{pmatrix} 2X-1 & X & X-1 & 1\\ X & 0 & 1 & 0 \\ 0 & 1 & X & X\\ 1 & X^2 & 0 & 2X-2 \end{pmatrix}$$

I know of two methods for computing the smith normal form of a matrix:

1) Using the algorithm

2) Computing the Fitting Ideals

The first process looks horrible and cumbersome for such a matrix- maybe it all cancels nicely but I have tried and it doesn't appear clean.

The second looks like it might work if the matix behaves 'nicely'- for example it isn't hard to see that $\mathrm{Fit}_1 (M)= \mathrm{Fit}_2 (M)= (1) $ so we have the first two diagonal elements as 1. However computing the 3x3 sub determinants seems really lengthy and can't see any nice terms like 1 coming out. I could compute the whole 4x4, but again, this seems lengthy and then I'd hope it to not have many factors so I can 'deduce' the 3rd diagonal element, but this seems too hopeful.

Is there some property of the ring/ overall technique I'm missing?

Any help would be appreciated

$\endgroup$
2
$\begingroup$

I don't think using row and column operations is so bad if you have a computer, at least. (Assuming I didn't make a mistake, of course.) I computed the Smith normal form using Sage using the following code:

R.<x> = PolynomialRing(QQ,1)
M = matrix(R,[[2*x-1, x, x-1, 1],[x, 0, 1, 0],[0, 1, x, x],[1, x^2, 0, 2*x-2]])
M.swap_rows(0,3)
M.add_multiple_of_row(1,0,-x)
M.add_multiple_of_row(3,0,-(2*x-1))
M.add_multiple_of_column(1,0,-x^2)
M.add_multiple_of_column(3,0,-(2*x-2))
M.swap_rows(1,2)
M.add_multiple_of_column(2,1,-x)
M.add_multiple_of_column(3,1,-x)
M.add_multiple_of_row(2,1,x^3)
M.add_multiple_of_row(3,1,-(-2*x^3+x^2+x))
M.add_multiple_of_column(3,2,-1)
M.add_multiple_of_row(3,2,-2)
M.swap_rows(2,3)
M.swap_columns(2,3)
M.add_multiple_of_row(3,2,2*x)
M.add_multiple_of_row(3,2,-6)
M.rescale_row(3,-1/13)
M.swap_rows(2,3)
M.add_multiple_of_row(3,2,-(x+2))
M.add_multiple_of_column(3,2,-(1/13*x^4 - 4/13*x^3 - 8/13*x^2 + 12/13*x - 19/13))
M.rescale_row(3,-13)
M

In the end, I get $$ \left(\begin{array}{rrrr} 1 & 0 & 0 & 0 \\ 0 & 1 & 0 & 0 \\ 0 & 0 & 1 & 0 \\ 0 & 0 & 0 & x^{5} - 2 x^{4} - 3 x^{3} + 9 x^{2} - 8 x + 1 \end{array}\right) \, . $$

You can try the code yourself online using the SageMathCell.

$\endgroup$
1
$\begingroup$

Since you've done the $1\times 1$ and $2\times 2$ minors ....

The $3\times 3$ minors got by deleting the first row and first column, and first row and last column are pretty obviously $X^3-2X+2$ and $X^5+1$ which are coprime.

Now take a deep breath and calculate the determinant: it's $X^5-2X^4-3X^3+9X^2-8X+1$. (Thanks to @Quasicoherent.)

But I think that it's usually best to grit one's teeth and do the Gauss eliminations.

$\endgroup$
  • $\begingroup$ Thanks, but should that be $X^4 +1$ for the second determinant? $\endgroup$ – Evgeny T Apr 24 '17 at 8:31
  • 1
    $\begingroup$ You are right and it is wise to check my sums! $\endgroup$ – ancientmathematician Apr 24 '17 at 9:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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