6
$\begingroup$

Clearly $G/H$ is cyclic if $G$ is cyclic. But sometimes $G/H$ is cyclic without $G$ necessarily cyclic, for example: $(\mathbb{Z}_6\times \mathbb{Z}_6)/\left \langle (2,3) \right \rangle \cong \mathbb{Z}_6$

In cases like the latter, how can one determine if the group is cyclic? I mean, is there something in general that lets me know?

For example

Write $(\mathbb{Z}_{20}\times \mathbb{Z}_6)/\left \langle (10,2) \right \rangle $ as an external direct product of cyclic groups of prime power order.

Is $(\mathbb{Z}_{20}\times \mathbb{Z}_6)/\left \langle (10,2) \right \rangle $ isomorphic to $\mathbb{Z}_{4}\times\mathbb{Z}_{5}$ or isomorphic to $\mathbb{Z}_{2}\times \mathbb{Z}_2\times \mathbb{Z}_5$?

$\endgroup$
4
  • 1
    $\begingroup$ Do you assume G to be abelian or not necessarily? $\endgroup$
    – Keen
    Mar 25 at 22:54
  • $\begingroup$ @Keen Well, I am working with Abelian groups, I missed $\endgroup$
    – Hopmaths
    Mar 25 at 22:55
  • $\begingroup$ You could try to calculate the order of $(0,1),(1,0)\in\mathbb Z_{20}\times\mathbb Z_6/\langle(10,2)\rangle$ $\endgroup$
    – Kenta S
    Mar 25 at 22:59
  • 2
    $\begingroup$ Smith normal form is often a useful tool for calculations of this type. In this case, you're looking at the cokernel of the matrix $\begin{bmatrix} 10 & 20 & 0 \\ 2 & 0 & 6 \end{bmatrix}$. $\endgroup$ Mar 25 at 23:58
3
$\begingroup$

According to the fundamental theorem of Abelian groups, you can write $G$ as a product $\mathbb{Z}_{n_1} \times \cdots \times \mathbb{Z}_{n_m}$. Now consider $f$ the morphism from $\mathbb{Z}^m$ to $G$, that consists of natural projections onto $n_i$. Let $\pi$ be the natural projection of $G$ onto $\frac{G}{H}$. Then $\frac{G}{H}$ is isomorphic to $\frac{\mathbb{Z}^m}{ker(\pi \circ f)}$.

Now we use the structure theorem of finitely generated modules over a PID. There exists a basis $(e_1 ,\cdots ,e_m)$ of $\mathbb{Z}^m$ and unique sequence of positive $d_1$ dividing $d_2$ dividing $d_3$ etcetera dividing $d_r$, such that $(d_1 e_1 , \cdots ,d_r e_r)$ is a basis of $ker(\pi \circ f)$, with $r$ being the rank of $ker(\pi \circ f)$. From this, we can simply deduce that $\frac{\mathbb{Z}^m}{ker(\pi \circ f)}$ is cyclic if and only if: $m=r$ and $d_1=\cdots = d_{m-1}=1$.

These $d_i$ can be calculated algorithmically if one knows a generating family of $ker(\pi \circ f)$. Now if we take $u_1, \cdots , u_k$ a family in $G$ generating $H$ (which is how one would typically describe a subgroup). Then if you take $v_1, \cdots , v_k \in \mathbb{Z}^m$, such that $f(v_i)=u_i$, then $v_1 , \cdots, v_m$ together with a family generating $ker(f)$ generate $ker(\pi \circ f)$.

Using that, you calculate the invariants $d_1, \cdots, d_m$ for the submodule $ker(\pi \circ f)$ and you can then conclude on whether $\frac{G}{H}$ is cyclic.

Now I am sure that for most cases you'll encounter, this method is an overkill, but it should work in general.

$\endgroup$
3
$\begingroup$

If $|G/H|$ is prime, then of course it's cyclic.

In general, the Smith normal form can be used to compute the structure of quotients of $\mathbb Z^n $. But every finitely generated abelian group is such a quotient. After determining $G/H$ as a quotient of the free abelian group on $n$ generators, this gives an approach.

$\endgroup$

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.