Show that $G=HK.$

abelian-groupsgroup-theoryproof-verificationproof-writing

Let $G$ be an abelian group of order $mn$ where $\text {gcd}\ (m,n)=1.$ Consider the subgroups $$\begin{align*} H & = \left \{g \in G : \text {Ord}\ g \ |\ m \right \};\\ K & = \left \{g \in G : \text {Ord}\ g \ |\ n \right \}; \end{align*}$$ of $G.$ Show that $G=HK.$

My attempt $:$ Let us take the prime factorizations of $m$ and $n.$ $$\begin {align*} m & = {p_1}^{{\alpha}_1}{p_2}^{{\alpha}_2} \cdots {p_r}^{{\alpha}_r};\\ n & = {q_1}^{{\beta}_1}{q_2}^{{\beta}_2} \cdots {q_s}^{{\beta}_s}; \end{align*}$$ where $p_1,p_2,\cdots ,p_r;q_1,q_2, \cdots , q_s$ are distinct primes and ${\alpha}_i \geq 1,; \beta_j \geq 1$ for $i=1,2,\cdots,r;j=1,2, \cdots , s.$ Now let $g \in G.$ By Lagrange's theorem $\text {Ord}\ g\ |\ mn.$ So $$\text {Ord}\ g = {p_1}^{{\gamma}_1}{p_2}^{{\gamma}_2} \cdots {p_r}^{{\gamma}_r} {q_1}^{{\delta}_1}{q_2}^{{\delta}_2} \cdots {q_s}^{{\delta}_s}$$ where ${\gamma}_i \geq 1; {\delta}_j \geq 1$ for $i=1,2, \cdots ,r; j=1,2, \cdots ,s.$ Let $r = {p_1}^{{\gamma}_1}{p_2}^{{\gamma}_2} \cdots {p_r}^{{\gamma}_r}$ and $s = {q_1}^{{\delta}_1}{q_2}^{{\delta}_2} \cdots {q_s}^{{\delta}_s}.$ Clearly $r\ |\ m$ and $s\ |\ n.$ Also $\text {Ord}\ g^s = r$ and $\text {Ord}\ g^r = s.$ Hence $g^r \in H; g^s \in K.$

Claim $:$ $\text {gcd}\ (r,s) = 1.$

Since $\text {gcd}\ (r,s)\ |\ r$ and $\text {gcd}\ (r,s)\ |\ s$ so it follows that $\text {gcd}\ (r,s)$ is a common divisor of $m$ and $n.$ Hence $\text {gcd}\ (r,s)\ |\ \text {gcd}\ (m,n) = 1.$ This proves our claim.

Since $\text {gcd}\ (r,s) = 1$ so by Bezout's theorem $\exists$ $a,b \in \Bbb Z$ such that $ar+bs = 1.$

Let $h = g^{ar} = \left (g^r \right )^a \in H$ and $k = g^{bs} = \left (g^s \right )^b \in K.$

Then $g = g^1 = g^{ar+bs} = g^{ar} g^{bs} = hk \in HK.$ So $G \subseteq HK \subseteq G.$ Therefore it follows that $G=HK.$ This completes the proof.

Is there any easy way to proof this result? Any suggestion regarding this will be highly appreciated.

Thank you very much.

EDIT $:$ I have just seen that in my entire proof I haven't used abelian condition anywhere. Is it of no importance here?

Best Answer

Define $\phi: G \to G$ as $\phi(g)=g^m$. Then because $G$ is abelian, $\phi$ is a homomorphism. Note that $\ker(\phi) = H$ and $k \to Hk$ defines an isomorphism from $K \to G/H$ so $G/H \cong K$.

Related Question