第13篇

\section{On a Class of Elementary Symmetric Functions}%13
\markboth{Articles}{On a Class of Elementary Symmetric Functions}

\vspace{4.2cm}

\subsection{Introduction}
It is a well-known and classic result that if p is a prime, then we have:
xp11(x1)(x2)(x(p1))(modp).

Hence the remainder of the elementary symmetric functions of the numbers
1,2,, p1(modp)
are zero, except for
σp1.
We can obtain that their product is 1(modp) which is Wilson's theorem.
Moreover, Wolstenholme proved that the numerator of the fraction
\ds\f11+\ds\f12+\ds\f13++\ds\f1p1
when written in lowest terms is a multiple of p2 for every prime number p5.

Despite Wolstenholme's result, this type of problems keeps appearing even
in olympiads.
For example in the XX Iberoamerican Olympiad held in Cartagena,
Colombia, one of the problems was to prove that the numerator of the
fraction
\ds\f11p+\ds\f12p+\ds\f13p++\ds\f1(p1)p,
when written in lowest terms, was divisible by p3 for every p5.

The purpose of this article is not only to prove these results, but also to
provide a solution for a more general case, showing some ideas that might be
useful for this kind of problems.
Of course all the discussed problems have
elementary solutions; however, the techniques used for them do not work in a
more general case, and we are forced to use stronger artillery.

It is time to do some mathematics before the reader gets bored of all this
talking.

\subsection{The Result}
Before continuing, let's first introduce the theorem we are going to prove.


{\bf Theorem 1.}
{\it Let p be a prime number greater than 3 and k a positive integer.
Let
P(x)=(x1pk)(x2pk)(x3pk)(x(p1)pk)=xp1xp2A(1,k)+xp3A(2,k)xA(p2,k)+A(p1,k).

Then the following statements are true:

1. A(i,k) is divisible by pk+1 for 1ip3.

2. A(p2,k) is divisible by pk+2.

3. A(p1,k)+1 is divisible by pk+1.
}

Before proving this result we must prove the following lemma.

{\bf Lemma 1.}
(Lifting the exponent).
{\it Let p be an odd prime.
If ab(modpt) then apbp(modpt+1).
}

{\bf Proof.}
Let a=b+mpt, therefore
ap=bp+k=1p\ds(pk)bpkmkptk.

Looking at this relation (modpt+1) and since
p\ds(pk) for k=1,2,,p1,
it follows that in fact apbp(modpt+1), and this completes the proof of the
lemma.

Now we can continue, but for those who thought we were only going to use
classical number theory arguments, you were wrong!
Instead, we try a more
algebraic approach by considering the polynomial
Q(x,y)=(xy)(xy2)(xyp1)=xp+xp2f1(y)+xp3f2(y)++xfp2(y)+fp1(y).

We will have to work a little with this polynomial, and the reason we picked it
will be clear later.
The first thing we are going to prove is that ϕp1(x) divides
fi(x) for i=1,2,,p2 in Z[x], where ϕp1(x)
is the (p1)th
cyclotomic polynomial.

This follows easily by plugging y=ζ where ζ is a (p1)th primitive root of
unity.
For this value we get
Q(x,ζ)=xp11
so
fi(ζ)=0 for i=1,2,,p2
because the latter is a polynomial identity in x.
Since the roots of ϕp1(x)
are precisely the (p1)th primitive roots of unity, and
ϕp1(x)Z[x],
we get the desired divisibility.

Now it is known that a primitive root mod p2 is also a primitive root mod
pt for every positive integer t.
Let g be a primitive root mod pt for all t (just pick one mod p2 and it works),
then
pk+1ϕp1(gpk).

This is not hard at all, since
pk+1(gpk)p11,
pk+1 must divide some cyclotomic polynomials that are divisors of xp11
evaluated at gpk.
Suppose
pϕd(gpk) then
pgpkd1
and since the order of g mod p is (p1) and dp1,
we must have d=p1,
proving that all factors of p must divide ϕp1(gpk),
and hence we get the desired divisibility.

Now, using the last result it is easy to finish the proof of the first part of the theorem.
The numbers g1pk,g2pk,,g(p1)pk
are in some order the remainders
1pk,2pk,3pk,,(p1)pk mod pk+1
because of the lifting lemma.
Now consider
Q(x,gpk)=(xgpk)(xg2pk(xg(p1)pk)=xp1+xp2f1(gpk)++xfp2(gpk)+fp1(gpk).

By the previous remarks we now have that
fi(gpk)0(modpk+1), for i=1,2,,p2, but
fi(gpk)A(i,k)(modpk+1),
also from the previous remarks, thus the first part of the theorem follows.

In order to prove the second part we must deal with a complicated divisibility
relation first; we prove it here as a lemma.

{\bf Lemma 2.}
{\it Let p be an odd prime and k a positive integer, then for every
integer a not divisible by p we have
p2k+3p2k+2pk+1(apk+(pa)pk).

}

{\bf Proof.}
The divisibility is equivalent to
pk+2pk+1(apk+(pa)pk)
and we will prove it by induction on k using the lifting lemma.
The cases k=0, k=1
are directly verified and are easily reduced to Fermat's little theorem.

The inductive step is not easy at all:
pk+1=apk+(pa)pk(modpk+2)
pk+1apk=(pa)pk(modpk+2)
(pk+1apk)=(pa)pk+1(modpk+3)
pk+2a(p1)pkapk+1=(pa)pk+1(modpk+3)
pk+2=apk+1+(pa)pk+1(modpk+3).

The third step is due to the binomial formula and the second one is by the
lifting lemma.
And that finishes the proof of the lemma.

To prove the second part of the theorem we will prove that the difference
P(pk+1)A(p1,k)
is divisible by p2k+3.
Multiplying and using Lemma 2,
(pk+1apk)(pk+1(pa)pk)apk(pa)pk(modp2k+3),
and multiplying this congruence for
a=1,2,,(p1)/2
we obtain the stated divisibility,
since
A(p1,k)=\dsi=1p1ipk.
We can also expand the difference as
P(pk+1)A(p1,k)=p(k+1)(p1)p(k+1)(p2)A(1,k)
+p(k+1)(p3)A(2,k)pk+1A(p2,k),
and since p2k+3 divides all terms but the last one in that sum (by the first
part of the theorem and a simple p-adic valuation), we get that p2k+3 must
divide the last term.
But this in turn implies that pk+2A(p2,k) as wished.

Note that Wolstenholme's theorem and the Iberoamerican Olympiad problem
follow from this part of the theorem with k=0, k=1 respectively.

To finish with the third part of the theorem we only have to take Wilson's
congruence and apply the lifting k times.
Doing this we get
1pk2pk(p1)pk(1)pk(modpk+1),
which is exactly the third part of the theorem,
and this concludes our proof.

\bigskip
\hfill
{\Large Pascual R. Mesa, Colombia}

posted on   Eufisky  阅读(196)  评论(0编辑  收藏  举报

编辑推荐:
· 开发者必知的日志记录最佳实践
· SQL Server 2025 AI相关能力初探
· Linux系列:如何用 C#调用 C方法造成内存泄露
· AI与.NET技术实操系列(二):开始使用ML.NET
· 记一次.NET内存居高不下排查解决与启示
阅读排行:
· Manus重磅发布:全球首款通用AI代理技术深度解析与实战指南
· 被坑几百块钱后,我竟然真的恢复了删除的微信聊天记录!
· 没有Manus邀请码?试试免邀请码的MGX或者开源的OpenManus吧
· 园子的第一款AI主题卫衣上架——"HELLO! HOW CAN I ASSIST YOU TODAY
· 【自荐】一款简洁、开源的在线白板工具 Drawnix
< 2025年3月 >
23 24 25 26 27 28 1
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31 1 2 3 4 5

导航

统计

点击右上角即可分享
微信分享提示