3

I'm stuck on how to write the notation formally and clearly for the set over all the product in the form of $\lambda_{i_1}\lambda_{i_2}\dots\lambda_{i_r}$, where each $i_j$ could be $1,2,\dots,$ or $n$, and $n$ is greater than $r$. Note that something like $\lambda_{1}\lambda_{1}\dots\lambda_{1}\lambda_{1}$ (repeating $\lambda_{1}$ $r$ times) or $\lambda_{1}\lambda_{1}\dots\lambda_{1}\lambda_{2}$ (repeating $\lambda_{1}$ $(r-1)$ times followed by a single $\lambda_2$) should also be included in the set. The multiplication is supposed to be commutative, so the order doesn't matter.

My current attempt: $$ \left\{\prod_{j=1}^{r}\lambda_{i_j}:i_j\in\{1,2,\dots,n\}\right\} $$

It doesn't seem clear enough. Does anyone have a better idea? I really appreciate it so much!

Thinkpad
  • 383
  • 1
  • 8
  • 3
    The notation you have proposed seems perfectly clear, but what does it have to do with your title? – LSpice Sep 24 '22 at 14:51
  • To give justice to the above, let's consider $\ \lambda_j\ $ as variables. I'll try to write more in an *Answer*. – Wlod AA Sep 25 '22 at 00:46
  • 1
    @LSpice Sorry, first of all, in the title it is supposed to be $r$ instead of $k$. This is because I feel like every monomial of degree $r$ (thanks for Wlod AA telling me this notion) is kind of like choosing some variable $r$ times from $n$ variables. Though, I know it is still different, as '$n$ choose $r$' usually requires the chosen $r$ variables are different. So I was not quite sure how to summarize my question in a single sentence, and I just had to use my intuition to make a title. – Thinkpad Sep 25 '22 at 07:52

3 Answers3

1

Basic notation:

  • $\ \mathbb Z_+\ :=\ \{\,n\in\mathbb Z:\ n\ge 0\,\}$
  • $\ \mathcal R$ -- an arbitrary unitary commutative ring, $\ 1\in\mathcal R;$
  • $\ V $ -- an arbitrary set (finite or infinite);   it should be interpreted as a set of variables, as a generalization of the OP's set $\ \{\lambda_1\,\ldots\,\lambda_n\}.$

Remark:   Please, do not "correct" my notation (neither above nor below).

OP, your notation is fine and clear to me. It stands for the set of the monomials of degree $r,\ $ in $\ n\ $ variables (a variable doesn't have to appear explicitly in a monomial).

Possibly, the given expression for the number of monomials is not clear to mathematicians. It is indeed awkward but for objective reasons of no fault of OT.

The mentioning of a (commutative) ring is not crucial, it just tells us that the variables are commutative. Thus, we can simply talk about abstract monomials in commuting variables.

One way or another, I'll simply add an equivalent formula, it should be useful in any event. Thus, let $\ \mathcal R\ $ be an arbitrary commutative ring, and we consider elements $\ \lambda_i\ $ to be variables that run over $\ \mathcal R.\ $ OP (@Thinkpad) has defined:

$$ \Lambda_0(n\ r)\ :=\ \left\{\prod_{j=1}^{r}\lambda_{i_j}:i_j\in\{1,2,\dots,n\}\right\} $$

(just name $\ \Lambda_0\ $ is mine). This looks visually simple but actually is not simple conceptually. I'll write something visually not so simple but structural.

Let $$ J(V\ r)\ :=\ \left\{\ j\in\mathbb Z_+^{\ V}: \ \sum j\,=\,r\,\right\} $$ Then the general expression$\ \Lambda\ $ is defined as follows: $$ \Lambda(V\ r)\ :=\ \left\{\,\prod j:\ j\in J(V\ r)\,\right\} $$

This finally is a clean definition of monomials of degree $\ r\ $ (in variable for $\ V$).

We see that:

$$ \Lambda_0(n\ r)\ =\ \Lambda(\{1\,\ldots\,n\}\,\ r) $$

Wlod AA
  • 2,124
  • 1
    Of course this is correct, but, in terms of encountering it in a paper, if this is a one-off definition then I'd much prefer to read the original notation, since it involves no auxiliary definitions. On the other hand I agree that it is worthwhile to know that the sort of notational technique you suggest is available. (I also note that your definition and the proposed one are different for non-commutative rings, although in this case that's irrelevant since the ring--or we could even just say group, or even less--is explicitly said to be commutative.) – LSpice Sep 25 '22 at 09:05
  • 1
    @LSpice, yes, we both agree that the original definition is simple. I just hope that my version adds to the understanding and may assist one in proving some theorems. – Wlod AA Sep 25 '22 at 11:21
1

We use the common notation $[n]:=\{1,2,\ldots,n\}$ and consider as index set the cartesian product \begin{align*} [n]^r=\{\left(x_1,x_2,\ldots,x_r\right)\big|x_k\in[n], 1\leq k\leq r\} \end{align*}

Using this notation we can write the set of wanted products as \begin{align*} \color{blue}{\left\{\prod_{k=1}^r\pi_k(x)\Big|x\in[n]^r\right\}} \end{align*}

Here we use the projection operator $\pi_k$ to select the $k$-th component of an $r$-tuple: \begin{align*} \pi_k(x)=\pi_k(x_1,x_2,\ldots,x_r)=x_k \end{align*}

Markus Scheuer
  • 108,315
0

By the commutative property of multiplication (so that $\lambda_2 \cdot \lambda_1 \cdot \lambda_1 = \lambda_1 \cdot \lambda_2 \cdot \lambda_1 = \lambda_1 \cdot \lambda_1 \cdot \lambda_2$) your notation always works.
The only minor edit that I would perform, is to compactly specify something as:
Let $r,n \in \mathbb{Z}^+$, we have the set $$ \left\{\prod_{j=1}^{r}\lambda_{i_j}:i_j\in\{1,2,\dots,n\}, 1 \leq r < n \right\}\qquad $$ $$\qquad =\ \bigcup_{r=1}^{n-1} \left\{\prod_{j=1}^{r}\lambda_{i_j}:i_j\in\{1,2,\dots,n\}\right\} $$

This is fine (IMHO), knowing in advance that $\{\lambda_{i_1} \cdot \lambda_{i_2} \cdots \lambda_{i_{n-1}} \cdot \lambda_{i_n} \}$ ($n$ factors) can never belong to the above mentioned set (since $r<n$ by hypothesis).

Marco Ripà
  • 1,062
  • 2
  • 19
  • You and @Thinkpad (the OP) have defined different objects. What you have done is not "editing" but replacing one definition by a different definition that is not equivalent to the original definition. It's like claiming that you "know better". – Wlod AA Sep 24 '22 at 20:21
  • I don't think I quite agree with @WlodAA's comment that you have defined a different object, but this is at best an extremely confusing definition. Given your explicitly binding $r$ and $n$ outside the definition of the set, including $1 \le r < n$ as part of the set specifier doesn't actually change the set (except that it makes the set empty if $r \ge n$). If you had not explicitly bound $r$—and it's easy for a reader to miss it—then your set would include all products of length strictly less than $n$. – LSpice Sep 24 '22 at 20:41
  • @LSpice, now, after my EDIT, this should be explicit (hence clear). – Wlod AA Sep 24 '22 at 22:56
  • @WlodAA, re, it seems that you have edited the post to replace one definition by another—though am I imagining it? No edit history shows. (Maybe it got lost in the migration?) My point is that the post as written binds both $r$ and $n$ before writing the definition, so that $r$ cannot be regarded as free in the definition of the set any more than $n$ can—we would not, for example, regard the definition of the set as $\bigcup_{n = 1}^\infty \bigcup_{r = 1}^n {\ldots}$. The proposed definition is, I argue, just confusing, not wrong. – LSpice Sep 25 '22 at 00:08
  • The system MO+SE_Mathematics is buggy. I did my edit on MO but it didn't show that the EDIT was mine on SE-Mathematics here. – Wlod AA Sep 25 '22 at 00:35
  • @LSpice, to me the OP's formal formula is more binding than a messy (confusing) OP's intro to the formula. ###### PS. I didn't introduce any new formula (yet); I only wrote an equivalent form of Marco's formula. – Wlod AA Sep 25 '22 at 00:49
  • Since I cannot see errors in my past definition, but it is (IMHO) just a matter of different tastes/personal preferences, I'll delete my answer soon... the post has been modified and we already have an accepted answer. So, I don't think it is useful to keep this too. – Marco Ripà Sep 25 '22 at 08:26
  • The error (which I would rather call a confusion) is in putting "$1 \le r < n$" inside the set braces, when it is really a side condition. It seems to me rather like writing, to pick a silly example, "${2x : x \in \mathbb Z,\ 2 = 1 + 1}$" when one could write, if at all, "${2x : x \in \mathbb Z}$, $2 = 1 + 1$". If you do not agree with changes made to your post, then you should feel free to roll back to your original. – LSpice Sep 25 '22 at 08:49
  • No problem leaving the current version (avoid to delete it) if you think it can be useful for the forum readers. BTW, I've sometime seen the notation I used, putting a condition inside the ${\dots }$, even if this could be just a casual format. – Marco Ripà Sep 26 '22 at 09:46