I'm trying to prove that $\operatorname{lcm}(n,m) = nm/\gcd(n,m)$ I showed that both $n,m$ divides $nm/\gcd(n,m)$ but I can't prove that it is the smallest number. Any help will be appreciated.
-
Have you tried using the Fundamental Theorem of Arithmetic? (I.e. every natural number has a factorization into powers of primes that is unique up to re-ordering) – sharris Jun 09 '15 at 18:11
-
From what sharris said, it's a fairly standard proof for $m\cdot n = gcd(m,n) \cdot lcm(m,n)$ (which is equivalent to what you want to show). – anak Jun 09 '15 at 18:13
5 Answers
${\bf\small Proof}\ \, d\,| \gcd(m,n)\!\!\!\overset{\ \ \rm\color{darkorange}U}\iff d\,|\, m,n\!\!\overset{\rm\color{#0a0}{\rm R}\!\!}\iff $ $ m,n\,\raise -2pt{\Large |}\, \color{#c00}{\frac{mn}d}\!\!\overset{\ \ \rm\color{darkorange}U}\iff {\rm lcm}(m,n)\,\raise -2pt{\Large |}\, \frac{mn}d\!\!\overset{\rm\color{#0a0}{\rm R}\!\!}\iff d\,\raise -2pt{\Large|}\,\frac{mn}{{\rm lcm}(m,n)}$ where above we used: $\,\rm \color{darkorange}U = $ LCM & GCD Definition / Universal Property. $\,\ \small\bf QED$
More conceptually: elevating the key $\rm\color{#0a0}{R =}$ cofactor reflection symmetry yields a simpler proof: note $\,d\,\color{#0a0}\mapsto\, \color{#c00}{mn/d}\:$ bijects common divisors $\,d\,$ of $\,m,n\,$ with common multiples (that divide $\,mn).$ Being order-$\rm\color{#c00}{reversing}$, it maps the $\rm\color{#c00}{Greatest}$ common divisor to the $\rm\color{#c00}{Least}$ common multiple, i.e. $\,{\rm\color{#c00}{G}CD}(m,n)\,\color{#0a0}\mapsto\, mn/{\rm GCD}(m,n) = {\rm \color{#c00}{L }CM}(m,n).\,\ \small\bf QED$
See these posts for more on this $\:\!\rm\color{#0a0}{involution\ (reflection)}$ symmetry at the heart of gcd, lcm duality.
Note $ $ For completeness, in $\,\Bbb N\!:\ (d\mid g\!\iff\! d\mid \ell)\,\Rightarrow\, g = l\,$ follows as here: $ $ both $\,g\,$ and $\,\ell\,$ have the same set of divisors $\,d\,$ so they have the same greatest divisor (themselves). Or, putting $\,d\in\{g,\,\ell\}\,$ in $\,d\mid g\iff d\mid \ell\,$ shows $\,g,\ell$ divide each other $\,g\mid\ell\mid g\,$ so $\,g\le \ell\le g\,$ so $\,\ell = g\,$ (in a general domain we deduce they are associates $\,g\approx \ell,\,$ hence equal up to a unit multiple).
- 282,220
-
-
@Andrew That's a very common notation for the gcd, but I'll edit it to avoid confusion. – Bill Dubuque Sep 15 '23 at 21:09
Hint: For any $a,b$ real numbers: $\min(a,b)+\max(a,b)=a+b$.
Now, if we have $a=a_1^{p_1} a_2^{p_2}\ldots$ and similarly with $b$, if you use the equation I just mentioned for all $p_i$, you will get, that $\gcd(a,b)\cdot\operatorname{lcm}(a,b)=ab$.
- 3,512
Let's just do this directly. Let $g = \gcd(m,n)$. We need to prove that $\operatorname{lcm}(m,n) = \dfrac{mn}{g}$.
STEP $0$. (Preliminary stuff.)
DEFINITION $1$. $L = \operatorname{lcm}(m,n)$ if and only if
1. L is a multiple of m and of n.
2. If C is a multiple of m and of n, then C is a multiple of L.
LEMMA $2$. If $\gcd(a,b) = 1$ and $a \mid bc$, then $a \mid c$.
PROOF. If $\gcd(a,b) = 1$, then there exists integers $A$ and $B$ such that $aA + bB = 1$. It follows that $acA + bcB = c$. Since $a | acA$ and $a \mid bcB$, then $a \mid c$.
STEP $1$. $\dfrac{mn}{g}$ is a common multiple of $m$ and of $n$.
This is true because $\dfrac m g$ and $\dfrac n g$ are integers and $\dfrac{mn}{g} = \dfrac{m}{g}n = m \dfrac{n}{g}$.
STEP $2$. If $G$ is a common multiple of $m$ and of $n$, then $G$ is a multiple of $\dfrac{mn}{g}$.
Suppose $G = mM = nN$ for some integers $M$ and $N$. Then $\dfrac G g = \dfrac m g M = \dfrac n g N$.
Since $\gcd\left( \dfrac m g, \dfrac n g \right) = 1$, and $\dfrac m g M = \dfrac n g N$, then, by LEMMA $2$, $\dfrac m g \mid N$, say $N = \dfrac m g N'$ for some integer $N'$.
So $\dfrac{G}{g} = \dfrac{n}{g} N = \dfrac{m}{g} \dfrac{n}{g} N'$. It follows that $G = \dfrac{mn}{g} N'$ and so $G$ is a multiple of $\dfrac{mn}{g}$.
From STEP $1$, STEP $2$, and DEFINITION $0$, we can conclude that $\operatorname{lcm}(m,n) = \dfrac{mn}{\gcd(m,n)}$.
- 27,619
$\DeclareMathOperator{\lcm}{lcm}$Here is one way without using the Fundmental theorem of arithmetic just using the definitions
The definition of $lcm(a,b)$ is as follows:
$t$ is the lowest common multiple of $a$ and $b$ if it satisfies the following:
i) $a \mid t$ and $b \mid t$
ii) If $a \mid c$ and $b \mid c$, then $t \mid c$.
Similiarly for the $\gcd(a,b)$.
Here is my proof:
Case I: $\gcd(a,b) \neq 1$
Suppose $\gcd(a,b) = d$.
Then $ab = dq_1b = dbq_1 = d\cdot (dq_1q_2)$
Claim: $\lcm(a,b) = dq_1q_2$
$a = dq_1 \mid dq_1q_2$
$b = dq_2 \mid dq_2q_1$.
Supppose $\lcm(a,b) = c$. Hence $c \leq dq_1q_2$ .
To get the other inequality we have $dq_1|a$ and $dq_2|b$. Hence $dq_1 \leq a \leq c \leq dq_1q_2$. Similiarly for $dq_2$.
Suppose that $c$ is strictly less than $dq_1q_2$, so we have $dq_1q_2 < cq_2$ and $dq_1q_2 < cq_1$.
So $dq_1q_2 < c < cq_2 < dq_2^2q_1$ and $dq_1q_2 < c < cq_2 < dq_1^2q_2$, but $dq_1^2q_2 > dq_1q_2$ so $c < dq_1q_2$ and $c > dq_1q_2$ contradiction. Hence $c = dq_1q_2$.
Notice that the case where $\gcd(a,b) = 1$ we can just set $q_1 = a$ and $q_2 = b$, and the proof will be the same.
- 35,369
Here is a proof that works only in $\def\Z{\Bbb Z}\Z$, while the result is valid much more generally (basically in any rings where $\gcd$ and $\def\lcm{\operatorname{lcm}}\lcm$ are well defined). So it certainly will seem inferior to other approaches to certain tastes, but it has a hands-on quality that may please others.
Let $n,m$ be two positive integers. Consider the list of remainders modulo$~n$ of the multiples $0,m,2m,\ldots$ of$~m$ until some multiple $km$ has a remainder that is already present in the list (this remainder is not added). The term in the list to which it was equal can only be the initial$~0$, since otherwise the the two terms in the list at the positions just before the two that were found to be equal would also have to be equal to each other, which would contradict that this was the first occurrence of such an equality. Having a remainder$~0$, this $km$ is a multiple of$~n$, so by definition the least common multiple of $m$ and $n$.
Now the set of classes represented by the numbers of the list forms a subgroup of the additive group $\Z/n\Z$ (the class of $0$ is among them, and the set is closed under taking differences in $\Z/n\Z$). If the initial$~0$ is the only number on the list, namely if $n$ divides $m$, then the result to prove is obvious (since $\gcd(n,m)=n$ and $\lcm(n,m)=m$), Having dealt with this case, we can assume there is a nonzero remainder on the list; let $d>0$ be the smallest such remainder. By the mentioned subgroup property, $d$ divides $n$, and the numbers on the list are all multiples of $d$. In particular $d$ divides the remainder modulo$~n$ of $m$ (the second number on the list) and therefore$~m$ itself: $d$ is a common divisor of $n$ and $m$. Also $d$ was obtained as remainder of some multiple $sm$ of$~m$, so $d=sm-tn$ for some integers $s,t$ which shows that all common divisors of $n,m$ divide $d$ (since they divide both $sm$ and $tm$), which makes $d$ the greatest common divisor of $n$ and $m$.
Now it remains to count the numbers on the list in two ways. On one hand we produced $k$ terms, where $km=\lcm(n,m)$, so the number is $k=\lcm(n,m)/m$. On the other hand the list consists (in some permutation) of all multiples $ld$ of$~d$ with $0\leq{ld}<n$, so the number is $n/d$. Using $d=\gcd(n,m)$, we get $$ \frac{\lcm(n,m)}m = \frac{n}{\gcd(n,m)} \qquad\hbox{so}\qquad \lcm(n,m) = \frac{nm}{\gcd(n,m)}. $$ An example can clarify the construction: for $n=24$ and $m=110\equiv14\pmod{24}$, the list of remainders is $$ 0, 14, 4, 18, 8, 22, 12, 2, 16, 6, 20, 10, $$ with $12m=1320=\lcm(n,m)$ producing a remainder$~0$ at what would be the next position in the list, and of which $2=7m-32n=\gcd(n,m)$ is the smallest positive remainder. The equality comes from equating $12=24/2$.
- 119,547
-
-
@BillDubuque: Hi, nice to hear from you again. Doing as suggested, I find that proof has more colours, more formulas, way less words, and is about a related but different statement. There is no sequence of remainders, or Bézout relation there. With "hands on" I mean that doing a concrete example helps explain what is going on, and follow the argument; I don't think that applies to the proof you mention. Altogether a proof in a rather opposite style, though the final equation does occur there (in an abbreviated form). – Marc van Leeuwen Apr 13 '25 at 04:33
-
Attempting to extract from that proof one that would apply here, I get that the condition $j\in\frac{n}{(n,m)}\Bbb Z$ is equivalent to $$ \frac{n}{(n,m)}\mid j \iff n\mid(n,m)j\iff n\mid(nj,mj)\iff n\mid mj \iff [n,m]\mid mj \iff \frac{[n,m]}m\mid j $$ so to $j\in\frac{[n,m]}n\Bbb Z$, and it follows that $|\frac{n}{(n,m)}|=|\frac{[n,m]}m|$. Which is slick, but not really my argument. – Marc van Leeuwen Apr 13 '25 at 04:52
-
My point is this: we both compute $,{\rm ord}_{:!n}(m) = |m\Bbb Z_n|,$ in two ways, via gcd and lcm. I use general gcd (or ideal) laws, which you essentially replace by common direct proofs of these laws in this special case. $\ \ $ – Bill Dubuque Apr 13 '25 at 09:01
-
Btw, since your goal seems to be to give a "concrete, hands on" approach, a nice way to enhance intuition at an elementary level is to employ graphical visual realizations of subgroups of cyclic groups using periodic Roulette curves (e.g. Star Polygons, Spirograph curves), see e.g. here and here. This works very nicely here. I've had success with such graphical presentations even at grade school level. – Bill Dubuque Apr 13 '25 at 16:48
-
@BillDubuque Thank you, the spirograph illustration of generating a subgroup of a cyclic group seems like a helpful visualisation. – Marc van Leeuwen Apr 16 '25 at 14:37