1

Let $$ R = \mathbb{Z}_m\times\cdots\times\mathbb{Z}_m = \times_{i=1}^\ell\mathbb{Z}_m $$ I'm trying to find the additive order of elements of $R$.

The additive order $\text{ord}(a)$ of $a\in\mathbb{Z}_m$ is the smallest integer $k\in\mathbb{N}$ such that $$ k\cdot a\equiv0\;\bmod m. $$ And resp. for $r\in R$ $$ \text{ord}(r)=\text{min}\{k\in\mathbb{N}:\forall i:k\cdot r_i\equiv0\;\bmod m\}. $$

So far I know that for $a\in\mathbb{Z}_m$ $$ \text{ord}(a) = \frac{\text{lcm}(a, m)}{a} $$ so I would expect $$ \text{ord}(r) = \text{lcm}\left(\frac{\text{lcm}(r_1, m)}{r_1},\ldots,\frac{\text{lcm}(r_\ell, m)}{r_\ell}\right) $$ for $r=(r_1,\ldots,r_\ell)\in R$.

But here I'm already stuck. Is there a way to simplify this any further?

2 Answers2

1

Since you're only concerned with additive order, we can disregard the ring structure and the question becomes that is the order of an element from the direct product of groups. In that context, we have that the order of an element from the direct product of groups is the lcm of the order of each coordinate, which is exactly as you suggest. This can be seen fairly easily:

Suppose $g= (g_1,...,g_\ell)$ is an element of some direct product of $\ell$ groups. Let $L$ be the lcm of the orders of $g_1,...,g_\ell$. Then clearly $$ L(g_1,...,g_\ell) = (Lg_1,...,Lg_\ell) = 1 $$ So the order of $g$ is divisible by $L$. Suppose $g$ has order $k$ strictly less than $L$, then $kg_i = 1$ for all $g_1,...,g_\ell$, but for at least one such $g_i$, $k$ is not a multiple of its order so we have contraction. Thus $g$ has order $L$ and we can apply the result you invoked to get the order of the element in this specific case.

Also this is pretty, minor, but usually the direct product of rings, groups, etc. when you have a set of them is denoted as $$ \prod_{i= 1}^\ell\mathbb{Z}_m $$ instead of the $\times_{i = 1}^\ell \mathbb{Z}_m$, which may help you find answers to this kind of question more easily in the future.

Noah Solomon
  • 1,922
  • I'm a bit confused by your first formula; why is it the multiplicative inverse on the right hand side rather than the additive inverse ($0$)? – striderhobbit Jan 15 '21 at 21:34
  • I was also so keen to change your notation to reduce some interference with my own notation, if that's alright. Feel free to roll back – striderhobbit Jan 15 '21 at 21:50
  • This is just because I am proving the statement in group theoretic language because you are only concerned with the additive order of an element and traditionally this is written multiplicatively. You can think of it as just addition and then replace $(1)$ with $(0)$, I was just trying to give you a slightly more general answer. (and I approved your edit :) ) – Noah Solomon Jan 15 '21 at 22:05
  • 1
    Alright thanks, then I'm following your proof now! – striderhobbit Jan 15 '21 at 22:16
1

Hint: $\quad\DeclareMathOperator{\lcm}{lcm}\dfrac{\lcm(r_i,m)}{r_i}=\dfrac{m}{\gcd(r_i,m)}$, so $$\lcm\biggl(\frac{\lcm(r_1,m)}{r_1},\dots,\frac{\lcm(r_\ell,m)}{r_\ell}\biggr)=\lcm\biggl(\frac{m}{\gcd(r_1,m)},\dots,\frac{m}{\gcd(r_\ell,m)}\biggr)=\frac{m}{\gcd(r_1,\dots,r_\ell,m)}. $$ Added: I used the formula $$\operatorname{lcm}\Bigl(\frac m r,\frac ns\Bigr)=\frac{\operatorname{lcm}(m,n)}{\gcd(r,s)} $$ (There's an analogous formula for the $\gcd$ of two rational numbers.)

Bernard
  • 179,256
  • Thanks, but I think the last equality may need further explication... – striderhobbit Jan 14 '21 at 05:59
  • This is because: $1)$ the numerators are the same, so we can factor out $m$, and thus we have only the l.c.m. of the reciprocals of the g.c.d.s, which the reciprocal of the g.c.d. thereof. $2)$ the g.c.d. is associative. Is it clearer? – Bernard Jan 14 '21 at 10:17
  • Ok, I was not thinking of taking the gcd of rationals before. I found that $\text{lcm}(m/a,m/b)=m/\text{gcd}(a,b)$ see here which after some more steps indeed yields your formula. – striderhobbit Jan 15 '21 at 21:27
  • Maybe it would be useful to add this formula to your answer? It's maybe not completely obvious – striderhobbit Jan 15 '21 at 21:30
  • 1
    You're probably right.. I thought it was known to everyone, but it seems some results in arithmetic are not so widespread nowadays. I'll add the general formula (when the numerator is not necessarily the same). – Bernard Jan 15 '21 at 22:40