Maths Olympiad Prep

Library / /371 of 397

Combinatorics Difficulty 7.1 National Olympiad, round 2 Prove it Taiwan

SS 為一個有限集合, 且令 A\mathcal{A} 為所有由 SS 映至 SS 之函數所成的集合。令 ffA\mathcal{A} 的一個元素且令 T=f(S)T = f(S) 表示 SSff 之下所對應之值域。假設對每一個 A\mathcal{A} 中的元素 gggfg \neq f, 皆滿足 fgfgfgf \circ g \circ f \neq g \circ f \circ g.
試證: f(T)=Tf(T) = T.

Let SS be a finite set, and let A\mathcal{A} be the set of all functions from SS to SS. Let ff be an element of A\mathcal{A} and let T=f(S)T = f(S) denote the image of SS under ff. Suppose that for every gg in A\mathcal{A} with gfg \neq f, we have fgfgfgf \circ g \circ f \neq g \circ f \circ g.
Prove that f(T)=Tf(T) = T.

Solution

For n1n \ge 1, denote the nn-th composition of ff with itself by
fn=deffff. f^n \stackrel{\text{def}}{=} f \circ f \circ \dots \circ f.
By hypothesis, if gAg \in \mathcal{A} satisfies fgf=gfgf \circ g \circ f = g \circ f \circ g, then g=fg = f. A natural idea is to try to plug in g=fng = f^n for some nn in the expression fgf=gfgf \circ g \circ f = g \circ f \circ g in order to get fn=ff^n = f, which solves the problem:

Claim: If there exists n3n \ge 3 such that fn+2=f2n+1f^{n+2} = f^{2n+1}, then the restriction f:TTf: T \to T of ff to TT is a bijection.

Proof. Indeed, by hypothesis,
fn+2=f2n+1ffnf=fnffnfn=f. f^{n+2} = f^{2n+1} \Leftrightarrow f \circ f^n \circ f = f^n \circ f \circ f^n \Rightarrow f^n = f.
Since n21n-2 \ge 1, the image of fn2f^{n-2} is contained in T=f(S)T = f(S), hence fn2f^{n-2} restricts to a function fn2:TTf^{n-2}: T \to T. This is the inverse of f:TTf: T \to T. In fact, given tTt \in T, say t=f(s)t = f(s) with sSs \in S, we have
t=f(s)=fn(s)=fn2(f(t))=f(fn2(t)). t = f(s) = f^n(s) = f^{n-2}(f(t)) = f(f^{n-2}(t)).
i.e., fn2f=ffn2=id on Tf^{n-2} \circ f = f \circ f^{n-2} = \text{id on } T
(here id stands for the identity function). Hence, the restriction f:TTf: T \to T of ff to TT is bijective with inverse given by fn2:TTf^{n-2}: T \to T.

It remains to show that nn as in the claim exists. For that, define
Sm=deffm(S)(Sm is the image of fm). S_m \stackrel{\text{def}}{=} f^m(S) \quad (S_m \text{ is the image of } f^m).
Clearly the image of fm+1f^{m+1} is contained in the image of fmf^m, i.e., there is a descending chain of subsets of SS
SS1S2S3S1 S \supseteq S_1 \supseteq S_2 \supseteq S_3 \supseteq S_1 \supseteq \dots
which must eventually stabilise since SS is finite, i.e., there is a k1k \ge 1 such that
Sk=Sk+1=Sk+2=Sk+3==defS. S_k = S_{k+1} = S_{k+2} = S_{k+3} = \dots \stackrel{\text{def}}{=} S_\infty.
Hence ff restricts to a surjective function f:SSf : S_\infty \to S_\infty, which also bijective since SSS_\infty \subseteq S is finite. To sum up, f:SSf : S_\infty \to S_\infty is a permutation of the elements of the finite set SS_\infty, hence there exists an integer r1r \ge 1 such that fr=idf^r = \text{id} on SS_\infty (for example, we may choose r=Sr = |S_\infty|!). In other words,
fm+r=fm on S for all mk.(1) f^{m+r} = f^m \text{ on } S \text{ for all } m \ge k. \qquad (1)
Clearly, Eq. (1) also implies that fm+tr=fmf^{m+tr} = f^m for all integers t1t \ge 1 and mkm \ge k. So, to find nn as in the claim and finish the problem, it is enough to choose mm and tt in order to ensure that there exists n3n \ge 3 satisfying
{2n+1=m+trn+2=m{m=3+trn=m2. \left\{ \begin{array}{l} 2n + 1 = m + tr \\ n + 2 = m \end{array} \right. \Leftrightarrow \left\{ \begin{array}{l} m = 3 + tr \\ n = m - 2. \end{array} \right.
This can be clearly done by choosing mm large enough with m3(modr)m \equiv 3 \pmod r. For instance, we may take n=2kr+1n = 2kr + 1, so that
fn+2=f2kr+3=f4kr+3=f2n+1 f^{n+2} = f^{2kr+3} = f^{4kr+3} = f^{2n+1}
where the middle equality follows by Eq. (1) since 2kr+3k2kr + 3 \ge k.

Want a route through all this instead of an archive? The track puts 2,000 problems in a working order, from AMC 10 level to the IMO shortlist.

Source: MathNet, licensed CC-BY-4.0. Statement translated into English from zh; metadata (topic, difficulty) added by this project.