Scribd
Upload
129 views4 pages

Introduction To Hook Length Formula PDF

This document provides a proof of the Hook Length Formula using a technique from Bandlow [1]. The proof is by induction and uses the following steps: 1) It defines key terms like Young tableaux, type, and hook length. 2) It states the Hook Length Formula theorem and proves it by induction on n, the number of squares. 3) The inductive step uses properties of removing the bottom-right corner square from each block of the Young tableaux. 4) It applies an identity from another paper [2] and takes two key limits to complete the proof.

Uploaded by

michael scott
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
129 views4 pages

Introduction To Hook Length Formula PDF

This document provides a proof of the Hook Length Formula using a technique from Bandlow [1]. The proof is by induction and uses the following steps: 1) It defines key terms like Young tableaux, type, and hook length. 2) It states the Hook Length Formula theorem and proves it by induction on n, the number of squares. 3) The inductive step uses properties of removing the bottom-right corner square from each block of the Young tableaux. 4) It applies an identity from another paper [2] and takes two key limits to complete the proof.

Uploaded by

michael scott
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Introduction to Hook Length Formula

Mehtaab Sawhney∗
July 22, 2016

Abstract
USAMO 2016 Problem 2 provided controversy on the nature of Hook
Length Formula (HLF) and whether it was “elementary”. The attached
note gives a variant of a proof by Bandlow [1] regarding HLF and has been
streamlined so that it requires a mere two limits and shows a possible route
of how a contestant could have proved the formula even on contest day.
(Although this would arguably have been more remarkable than solving
all problems on the test itself.)

1 Introduction
For the sake of brevity we will only briefly define the terms necessary for the
proof. A Young Tableaux is a grid of left-justified squares such that the rows
have a weakly decreasing number of squares as one goes down. Suppose there
are n squares in the Young Tableau, then the Hook Length Formula counts the
number of permutations of 1 through n in the grid so that the numbers increase
as one goes from left to right on the rows and up to down on the columns. (Also
known as Standard Young Tableaux.)

Figure 1: Young Tableaux (Taken from https://www.wikipedia.org/)

Finally the type of a Young Tableaux is a list of the number of squares in each
row. This above diagram has type (5, 4, 1). Note that the type corresponding
directly to a partition of n, and we denote the shape as λ. The key thing to note
is that in any Standard Young Tableaux, n is at the end of one of the rows such
∗ Student at the University of Pennsylvania

1
that the row below is strictly shorter than the row above. Finally the hook
length in “Hook Length Formula” is the sum of the number of squares strictly
below or to right of a given square plus 1. In particular the box filled with a 2
in the above figure has hook length 5 while the square filled in with 6 has hook
length 2.

2 Elementary Proof of Hook Length’s Formula


Bandlow [1] gives an elementary proof of Hook Length Formula, and this note
mimics that proof with a less technical presentation.
Theorem 1 Let the number of Standard Young Tableaux with n squares and
shape λ, be Qλ . Then
n!
Qλ = Q
t∈λ L(t)

with L(t) denoting the hook length of squares t in the λ.


Proof: The proof is by induction on n. For n = 1 the statement is trivial.
Suppose that the theorem holds for all n < k, and now we prove it for k.
Suppose λ has the type (b1 + . . . + b` , . . . , b1 + . . . + b` , b1 + . . . + b`−1 , . . . , b1 +
. . . + b`−1 , . . . , b1 , . . . , b1 ) where there are ai rows with length b1 + . . . + bi . Let
the sth block be the as rows of length b1 + . . . + bs . The only possible position
for k is the bottom-right corner of one of these blocks. Note that after removing
one of these corners then the remainder of the squares also form a Standard
Young Tableaux. Finally let λs be λ without the bottom-right corner of the sth
block. Note that the hook length of only the squares in the same row or column
as the removed square are affected. Then it follows that
Q
L(t)
Rs = Q t∈λ
t∈λs L(t)
    
a1 + . . . + as−1 + b1 + . . . + bs a2 + . . . + as + b2 + . . . + bs as + bs−1 + bs
= ... as bs
a1 + . . . + as−1 + b2 + . . . + bs a2 + . . . + as−1 + b3 + . . . + bs as−1 + bs
   
as + as+1 + bs+1 as + as+1 + as+2 + bs+1 + bs+2 as + · · · + a` + bs+1 + . . . + bs
×
as + bs+1 as + as+1 + bs+1 + bs+2 as + . . . + a`−1 + bs+1 + ` + b`
by straightforward computation. The inductive step is equivalent to
`
X `
X
X({ai ,bi },`) = Rs = k = ai (b1 + . . . + bi ).
s=1 i=1

This identity will now be proved by induction on `. Note that ` = 1 is trivial


and then the key is to use
 
a1 + . . . + as + b1 + . . . + bs b1
=1+ .
a1 + . . . + as−1 + b2 + . . . + bs a1 + . . . + as−1 + b2 + . . . + bs

2
for s ≥ 2. Now it follows that

X({ai ,bi },`) − X({ai ,bi }/{a1 ,b1 },`−1)


   
a1 + a2 + b2 b2 + . . . + b` + a1 + . . . + a`
= b1 a1 ...
b2 + a1 b2 + . . . + b` + a1 + . . . + a`−1
    
b2 a 2 + a 3 + b3 a2 + . . . + a` + b3 + . . . + b`
+b1 a2 ...
a1 + b2 a 2 + b3 a2 + . . . + a`−1 + b3 + . . . + b`
..
.
   
a2 + . . . + a`−1 + b2 + . . . + b` b`
+b1 a` ... .
a1 + . . . + a`−1 + b2 + . . . + b` a`−1 + b`
Thus to prove the identity by induction it suffices to show that

X({ai ,bi },`) − X({ai ,bi }/{a1 ,b1 },`−1) = b1 (a1 + . . . + a` ).

Now denote s1 = 0 and st = a1 + . . . + at−1 + b2 + . . . + bt for t ≥ 2. The


inductive step is equivalent to
` ` `
X X Y ai
ai = ai 1+ .
i=1 i=1
sj − si
1≤j6=k≤n

This is a well-known identity from [2] and the following proof is almost identical
to the one given in the paper.
Lemma 1 For c ∈
/ {b1 . . . , bn } then
n n n
Y x + ak − bk Y c + ak − bk X ak (x − c) Y bk + aj − bj
= + .
x − bk c − bk (bk − c)(x − bk ) bk − bj
k=1 k=1 k=1 1≤j6=k≤n

Proof: Let
n
Y
f (x) = (x + ak − bk )
k=1

and
n
Y
w(x) = (x − c) (x − bk ).
k=1

Using Lagrange Interpolation of f (x) at c, b1 , . . . , bk then


n n
Y x − bk X f (bk )w(x)
f (x) = f (c) +
c − bk (x − bk )w0 (bk )
k=1 k=1

and using Y
f (bk ) = ak (bk + aj − bj )
1≤j6=k≤n

3
it follows that
n n n
Y x + ak − bk Y c + ak − bk X ak (x − c) Y bk + aj − bj
= + .
x − bk c − bk (bk − c)(x − bk ) bk − bj
k=1 k=1 k=1 1≤j6=k≤n

Therefore the lemma follows. The final trick is to take a pair of limits. Taking
the limit as c → ∞ it follows that
n   n Y  
Y ak X ak aj
1+ =1+ 1+ .
x − bk x − bk bk − bj
k=1 k=1 1≤j6=k≤n

Multiplying by x it follows that


n 
Y   Xn
ak ak x Y aj
x 1+ −1 = 1+ .
x − bk x − bk bk − bj
k=1 k=1 1≤j6=k≤n

Finally taking the limit as x → ∞ then


n n
X X Y aj
ak = ak 1+ .
i=1 i=1
bk − bj
1≤j6=k≤n

Shifting the index n to ` and taking the sequencs {ai } and {si } then
` ` `
X X Y ai
ai = ai 1+
i=1 i=1
sj − si
1≤j6=k≤n

as desired. Therefore the inductive step is complete and the result follows by
induction.

References
[1] Jason Bandlow, An elementary proof of the hook formula, Electron. J. Com-
bin 15 (2008), no. 1.

[2] R William Gosper, Mourad EH Ismail, Ruiming Zhang, et al., On some


strange summation formulas, Illinois Journal of Mathematics 37 (1993),
no. 2, 240–277.

You might also like