Author: Eiko

Tags: algebra, bourbaki, exercises

Time: 2025-01-10 06:45:40 - 2025-01-13 07:00:39 (UTC)

This is a documentation of some interesting exercises in Bourbaki.

Chapter 1

Section 6, Exercise 30

Let p be a prime and r,t non-negative integers, C is a cyclic group of order pr+t and TC a subgroup of order pt. So pr=[C:T]=|C/T|.

  • If C acts on a finite set E, we have

    |E||ET|modpr+1=p[C:T].

    think of ET as something that interpolate between E and EC, where |E|=|E| and |E||EC|modp.

  • As an application, we have

    (pr+tmptn)(prmn)modpr+1.

Proof

  • This is an application of orbit formula

    |E|=[C:Stab(e)]

    we have that eET iff TStab(e) iff [C:Stab(e)] is a factor of [C:T] (Note this iff requires cyclic!). So that otherwise, eET iff [C:Stab(e)] is divisible by p[C:T], thus the result.

  • Let [n] denote a finite set of size n. Consider the action of C on C×[m] given by acting on the first coordinate. This induces an action of C on E=(C×[m]|T|n), the set of subsets of C×[m] of size ptn=|T|n.

    The projection map φ:C×[m]C/T×[m] is C-equivariant and also T-invariant (though these facts are not used). We claim that this induces a bijection

    φ:ET=(C×[m]|T|n)T(C/T×[m]n).

    If a set FE is fixed by T, it is given a T-action on F. This action is free (all stablizers are trivial) so every orbit is of size |T|, mapping each orbit (there are n of them) T(s,x) to TsC/T gives the desired bijection. This shows

    |ET|=|(C/T×[m]n)|=(prmn).

    Applying the previous result, and use |E|=(pr+tmptn), we get

    (pr+tmptn)(prmn)modpr+1.

Inspired: A New Proof of Binomial Mod p Expansion Using Group Actions

The above proof inspires me to come up with a new proof of the binomial expansion mod p, i.e.

(nm)(nkmk)(n0m0)modp

where n=nkpk+nk1pk1++n0 and m=mkpk+mk1pk1++m0 with 0mi,ni<p.

Let Ci be the cyclic group of order pi. Consider the set

X=i=0kCi×[ni].

This set is endowed with a natural P=Ck××C0 action given by multiplying on the first coordinate. The action naturally induces an action on the set E=(Xm) of size m subsets of X.

Now let’s investigate what is EP. We have that a set FEP would imply a P action on F, for which the orbit sizes are computed as

|P(ci,x)|=[P:Stab((ci,x))]=|Ci|=pi.

This together with the uniqueness of p-adic expansions tells us that the size of F(Ci×[ni]) must be pimi, consisting of exactly mi orbits of size pi. Collapsing the orbits and mapping to the set i=0k([ni]mi) gives the bijection

EP=(Xm)P1:1i=0k([ni]mi).

Finally, utilizing the orbit formula, for any p-group |E||EP|modp, we get the desired result

(nm)(nkmk)(n0m0)modp.