2

If $M=\mathbb{Z}\oplus \cdots \oplus \mathbb{Z}$ is a finitely generated free $\mathbb{Z}$-module of rank $n$, and if we have $n$ vectors $v_1,\ldots,v_n\in M$ which are $\mathbb{Z}$-independent, then the subgroup $\mathbb{Z}v_1 + \cdots + \mathbb{Z}v_n$ is of finite index which can be computed using determinant (with absolute value) of $n\times n$ integer matrix, whose $n$ rows are $v_1,\ldots,v_n$.

Now consider the situation where module is over ring bigger than $\mathbb{Z}$:

More specifically, let $M=\mathbb{Z}[\omega]\oplus \mathbb{Z}[\omega]\oplus \mathbb{Z}[\omega]$ and consider following vectors in it:

[Edit: $\omega$ is primitive cube root of $1$ in $\mathbb{C}$.]

$$v_1=(1,1,1), \,\, v_2=(1,\omega,\omega^2), \,\, v_3=(1,\omega^2,\omega).$$ Let $N=\mathbb{Z}[\omega]v_1 + \mathbb{Z}[\omega]v_2 + \mathbb{Z}[\omega]v_3$.

Q.1 How can we decide if index of additive subgroup $N$ in $M$ s finite or not? and how can we compute if it is finite?

Q.2 What is the relation of $[M:N]$ with determinant of matrix with rows $v_1,v_2,v_3$?

Beginner
  • 10,836
  • 2
    Is $\omega$ a primitive 3rd root of unity? If so, then $\mathbb{Z}[\omega]$ is a PID, so you can use the structure theorem for finitely generated modules over a PID. It comes down to computing the Smith normal form of a matrix; see this post or this post, for instance. – Viktor Vaughn Nov 15 '19 at 05:20
  • Oh sorry, I forgot to mention it. Yes, it is cube root of unity. – Beginner Nov 15 '19 at 05:33
  • Can you suggest some reference for question 2? – Beginner Nov 15 '19 at 05:38
  • 1
    As $\Bbb{Z}[\omega]$ is a PID, any submodule $N$ of $M$ is also a free $\Bbb{Z}[\omega]$-module. A basis can be found via calculations involving the Smith normal form, as André explained. You then get the index $[M:N]$ from the absolute value of the (field) norm of the determinant of the change of bases matrix. Basically because Smith (or the so called stacked bases theorem) reduces it to the case of submodules of $\Bbb{Z}[\omega]$ (=ideals), where it is easy to show that the norm of the generator is all you need. – Jyrki Lahtonen Nov 15 '19 at 05:59
  • @Jyrki: I didn't get the statement "You then get the index [M:N] from the absolute value of the (field) norm of the determinant of the change of bases matrix.". Can you elaborate? Also, determinant of $3\times 3$-matrix above is $\pm 3\sqrt{-3}$ [Edited after comment by Jyrki], whereas, treating $N$ as $\mathbb{Z}$-submodule of $M$ (both of rank $6$), the determinant of $6\times 6$-matrix is $27$; I didn't get connection between these two determinants. – Beginner Nov 17 '19 at 12:14
  • The determinant cannot be $3\sqrt{3}$ because that is not an element of $\Bbb{Z}[\omega]$. Probably the determinant is an associate of $3\sqrt{-3}$. Then you are to take the absolute value of the norm of that. And it comes to $$N^{\Bbb{Q}(\omega)}_{\Bbb{Q}} (3\sqrt{-3})=27.$$ – Jyrki Lahtonen Nov 17 '19 at 12:18
  • Dear Jyrki, thank you very much for clarification. Can you suggest some reference for abstract theory behind this fact (that compute original determinant and take norm to get index of subgroup?) – Beginner Nov 17 '19 at 12:20

0 Answers0