Scribd
Upload
7 views5 pages

Notes For Lecture On 18 November 6

The document discusses symmetric polynomials, which remain unchanged under permutations of their variables, and their connection to representation theory and combinatorics. It introduces key symmetric polynomials such as elementary, complete, and power sum polynomials, along with their generating functions and properties. Additionally, it covers the representation of symmetric polynomials indexed by partitions and provides a theorem on the number of non-negative integer tuples with a specified sum.

Uploaded by

Jetha Lal
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
7 views5 pages

Notes For Lecture On 18 November 6

The document discusses symmetric polynomials, which remain unchanged under permutations of their variables, and their connection to representation theory and combinatorics. It introduces key symmetric polynomials such as elementary, complete, and power sum polynomials, along with their generating functions and properties. Additionally, it covers the representation of symmetric polynomials indexed by partitions and provides a theorem on the number of non-negative integer tuples with a specified sum.

Uploaded by

Jetha Lal
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/ 5

Enumerative Combinatorics 6:

Symmetric polynomials
Peter J. Cameron
Autumn 2013

A symmetric polynomial in n indeterminates is one which is unchanged


under any permutation of the indeterminates. The theory of symmetric poly-
nomials goes back to Newton, but more recently has been very closely con-
nected with the representation theory of the symmetric group, which we
glanced at in Lecture 3. I will just give a few simple results here. The best
reference is Ian Macdonald’s book Symmetric Functions and Hall Polynomi-
als.

5.1 Symmetric polynomials


Let x1 , . . . , xn be indeterminates. If π is a permutation of {1, . . . , n}, we
denote by iπ the image of i under π. Now a polynomial F (x1 , . . . , xn ) is a
symmetric polynomial if

F (x1π , . . . , xnπ ) = F (x1 , . . . , xn ) for all π ∈ Sn ,

where Sn is the symmetric group of degree n (the group of all polynomials


of degree n).
Any polynomial is a linear combination of monomials xa11 · · · xann , where
a1 , . . . , an are non-negative integers. The degree of this monomial is a1 +· · ·+
an . A polynomial is homogeneous of degree r if every monomial has degree
r. Any polynomial can be written as a sum of homogeneous polynomials of
degrees 1, 2, . . ..
In a homogeneous symmetric polynomial of degree r, the exponents in
any monomial form a partition of r into at most n parts; two monomials
which give rise to the same partition are equivalent under a permutation,

1
and so must have the same coefficient. Thus, the dimension of the space
of homogeneous symmetric polynomials of degree r is pn (r), the number of
partitions of r with at most n parts.
There are three especially important symmetric polynomials:

(a) The elementary symmetric polynomial er , which is the sum of all the
monomials consisting
  of products of r distinct indeterminates. Note
n
that there are monomials in the sum.
r
(b) The complete symmetric polynomial  hr , which
 is the sum of all the
n+r−1
monomials of degree r. There are terms in the sum: the
r
proof of this is given in the Appendix to these notes.
X n
(c) The power sum polynomial pr , which is simply xri .
i=1

For example, if n = 3 and r = 2,

(a) the elementary symmetric polynomial is x1 x2 + x2 x3 + x1 x3 ;


(b) the complete symmetric polynomial is x1 x2 + x2 x3 + x1 x3 + x21 + x22 + x23 ;
(c) the power sum polynomial is x21 + x22 + x23 .
   
n n+r−1
Note that er (1, . . . , n) = , hr (1, . . . , 1) = , and pr (1, . . . , 1) =
r r
n.
Also, the q-binomial theorem that we met in the last lecture shows that
 
2 n−1 r(r−1)/2 n
er (1, q, q , . . . , q ) = q ,
r q

and Heine’s formula shows that, similarly,


 
2 n−1 n+r−1
hr (1, q, q , . . . , q )= .
r q

5.2 Generating functions


The best-known occurrence of the elementary symmetric polynomials is the
connection with the roots of polynomials. (To avoid conflict with xi , the

2
variable in a polynomial is t in this section.) The coefficient of tn−r in a
polynomial of degree n is (−1)r er (a1 , . . . , an ), where a1 , . . . , an are the roots.
This is because the polynomial can be written as

(t − a1 )(t − a2 ) · · · (t − an ),

and the term in tn−r is formed by choosing t from n − r of the factors and
−ai from the remaining r.
Said otherwise, and putting xi = −1/ai , this says that the generating
function for the elementary symmetric polynomials is
n
X n
Y
r
E(t) = er (x1 , . . . , xn )t = (1 + xi t),
r=0 i=1

with the convention that e0 = 1.


In a similar way, the generating function for the complete symmetric
polynomials is
X n
Y
H(t) = r
hr (x1 , . . . , xn )t = (1 − xi t)−1 .
r≥0 i=1

We also take P (t) to be the generating function for the power sum polyno-
mials, with a shift:
X
P (t) = pr (x1 , . . . , xn )tr−1 .
r≥1

Now we see that H(t) = E(−t)−1 , so that


n
X
(−1)r 3r hn−r = 0 for n ≥ 1.
r=0

For P (t), we have

d d
H(t) = P (t)H(t), E(t) = P (−t)E(t).
dt dt

3
5.3 Functions indexed by partitions
We extend the definitions of symmetric polynomials as follows. Let λ =
(a1 , a2 , . . .) be a partition of r, a non-decreasing sequence of integers with
sum r. Then, if z denotes one of the symbols e, h or p, we define zλ to be the
product of zai over all the parts ai of λ; this is again a symmetric polynomial
of degree r. For example, if n = 3 and λ is the partition (2, 1) of 3, we have

eλ = (x1 x2 + x1 x3 + x2 x3 )(x1 + x2 + x3 ),
pλ = (x21 + x22 + x23 )(x1 + x2 + x3 ),
hλ = eλ + pλ .

We also define the basic polynomial mλ to be the sum of all monomials with
exponents a1 , a2 , . . .. In the above case,

mλ = x21 x2 + x21 x3 + x22 x1 + x22 x3 + x23 x1 + x23 x2 .

Theorem 5.1 If n ≥ r, and z is one of the symbols m, e, h, p, then any


symmetric polynomial of degree r can be written uniquely as a linear com-
bination of the polynomials zλ , as λ runs over all partitions. Moreover, in
all cases except z = p, if the polynomial has integer coefficients, then it is a
linear combination with integer coefficients.

So the polynomials er or hr , with r ≤ n, are generators of the ring of


symmetric polynomials in n variables with integer coefficients. For z = e,
this is a version of Newton’s Theorem on symmetric polynomials (which,
however, applies also to rational functions).

5.4 Appendix: Selections with repetition


Theorem
 5.2 The number of n-tuples of non-negative integers with sum r
n+r−1
is .
r

The claim about the number of monomials of degree r follows immediately


from this result, which should be contrasted with
 the fact that the number
n
of n-tuples of zeros and ones with sum r is .
r

4
Proof We can describe any such n-tuple in the following way. Take a line
of n + r − 1 boxes. Then choose n − 1 boxes, and place barriers in these
boxes. Let

(a) a1 be the number of empty boxes before the first barrier;


(b) a2 be the number of empty boxes between the first and second barriers;
(c) . . .
(d) an be the number of empty boxes after the last barrier.

Then a1 , . . . , an are non-negative integers with sum r. Conversely, given n


non-negative integers with sum r, we can represent it with n − 1 barriers in
n + r − 1 boxes: place the first barrier after a1 empty boxes, the second after
a2 further empty boxes, and so on.
So the required number of n-tuples is equal to the number of ways to
position n − 1 barriers in n + r − 1 boxes, which is
   
n+r−1 n+r−1
= ,
n−1 r

as required.

You might also like