So I know $|G|=p_1p_2p_3\cdots p_n$ but I don't know where to go from there. I'm having trouble figuring out how they both relate.
-
Well, what theorems do you have at hand? Maybe the structure theorem? – Pedro Oct 20 '14 at 02:27
-
1The Fundamental Theorem of Finite Abelian Groups – Jessy White Oct 20 '14 at 02:29
-
Well, then use that. – Pedro Oct 20 '14 at 02:29
-
& the theorem that states that every finite abelian group G is the direct product of p-groups – Jessy White Oct 20 '14 at 02:29
-
Write $G$ as direct product of $p$-groups, then write $m$ as a product of prime numbers... – Quang Hoang Oct 20 '14 at 02:34
-
(Better duplicates may exist. The question here is a Problem Statement Question, while the other question is better. Except it is a "check my proof" question, so doesn't quite fit - except, none of the answers actually check the proof!) – user1729 Jul 20 '21 at 13:53
3 Answers
Hint: Step 1: By the fundamental theorem of finite abelian groups, you know that $G\cong\Bbb Z/(p_1^{k_1})\times\cdots \Bbb Z/(p_r^{k_r})$ (note that $p_i$ and $p_j$ need not be distinct for $i\neq j$). Then if $\left|G\right| = m$, what is $m$ in terms of the $p_i^{k_i}$'s? (What is $\left|\Bbb Z/(p^r)\right|$? If $H$ and $K$ are finite abelian groups, what is $\left|H\times K\right|$?)
Step 2: Writing $m = \prod_{p\textrm{ prime}}p^{r_p}$, $n = \prod_{p\textrm{ prime}}p^{s_p}$, what can you say about the $s_p$'s in terms of the $r_p$'s? (Is one bounded by the other for all $p$, perhaps?) How does the factorization of $m$ into primes relate to the expression you found for the order of $G$ in step 1?
Step 3: Next, you want to use your knowledge of the subgroups of a cyclic group: in particular, what are the subgroups of $\Bbb Z/(N)$ for arbitrary $N\in\Bbb Z$ (and what are their orders)?
Step 4: Using this information, you'll be able to use the decomposition of $G$ given by the fundamental theorem of finite abelian groups to find subgroups $H_i$ of each factor $\Bbb Z/(p_i^{k_i})$ such that $\left|H_1\times\cdots\times H_r\right| = n$. Here you'll need to take a bit of care, as even though $\left|G\right| = p_1^{k_1}\cdots p_r^{k_r}$, $G$ need not be isomorphic to $\Bbb Z/(p_1^{k_1})\times\cdots\times\Bbb Z/(p_r^{k_r})$ - so you may find that you have a prime factor $q$ of $n$ for which there is no factor $\Bbb Z/(p_i^{k_i})$ of $G$ with $q$ dividing $p_i^{k_i}$ (Why is this relevant? Think about step 3.). However, this can be remedied using the observations you made in step $2$.
- 23,855
If $|G|={p_1}^{k_1}{p_2}^{k_2}....{p_n}^{k_n}$ and $ n={p_1}^{r_1}{p_2}^{r_2}....{p_n}^{r_n}$ where $n|m$, and $p_i's$ may be equal or distinct and as $G \cong \mathbb{Z_{p_1^{k_1}}} \times \mathbb{Z_{p_2^{k_2}}}......\mathbb{Z_{p_n^{k_n}}} $ pick, $a_1 \in \mathbb{Z_{p_1^{k_1}}} $ such that $|a_1|=p_1^{r_1}$, and similarly $a_2,a_3,,,,a_n$ etc i.e. $|a_i|=p_i^{r_i}$. Then $a_1.a_2.a_3...a_n$ is an element of order $n$ as $G$ is abelian.
- 11,677
-
-
1Careful, $G$ need not be isomorphic to $\Bbb Z_{p_1^{k_1}}\times\cdots\times\Bbb Z_{p_r^{k_r}}$: $\left|\Bbb Z/(2)\times\Bbb Z/(2)\right| = 4$, but $\Bbb Z/(2)\times\Bbb Z/(2)\not\cong\Bbb Z/(4)$. – Stahl Oct 20 '14 at 02:56
-
-
-
I see now you're not assuming that the $p_i$'s are distinct. (Personally, I would make that explicit for clarity's sake.) – Stahl Oct 20 '14 at 03:02
-
-
Right. My concern is that in the first line it appears that you try to factor $\left|G\right|$ into a product of distinct primes to powers, but this does not tell you the product of cyclic groups that $G$ is isomorphic to. – Stahl Oct 20 '14 at 03:08
-
Now you allow for the primes to be equal, but you're choosing an arbitrary factorization of $G$ at the beginning, without knowing what product of cyclics $G$ is isomorphic to. I think you would want to say $G\cong\cdots$ first, then deduce that $\left|G\right| = \cdots$. – Stahl Oct 20 '14 at 03:10
-
yeah, i mean first use FT of finite ab group and then use order – Bhaskar Vashishth Oct 20 '14 at 03:20
-
I'm so sorry to bother you, but can you explain in a few words why the a's multiplied together is an element rather then a group of order n? – Jessy White Oct 20 '14 at 03:29
-
well it is like suppose your group is $Z/(2) \times Z/(4) \times Z/(44) \times Z/(100)$ of order $m=35200$ and let $n=550=5 \times 11 \times 10= 11 \times 50$ then the the order of subgroup generated by $(0,0,4,2)$ does the job. – Bhaskar Vashishth Oct 20 '14 at 03:52
Apply induction on order of $G$. It is true for $m=1$.
Let $p$ be a prime dividing $n$ ,then $p$ divides $m$. So $G$ has an element $a$ of order p. Consider $K=\langle a \rangle$, thus $O(G/K)=m/p$, which is less than $m$.
Let $n=m_1p$. So by the induction hypothesis $G/K$ has a subgroup of order $ m_1 $ (say, $H/K$) where $H$ is a subgroup of $G$ containing $K$. Hence, $O(H)=m_1p=n$.
- 2,266
- 31,675
- 10
- 110
- 250