Lecture 6

Pólya’s Enumeration Formula.

Stirling cycle numbers. Stirling set numbers

Isabela Drămnesc UVT

Computer Science Department,

West University of Timişoara,
Romania

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 1 / 25

Counting in the presence of symmetries

Burnside’s Lemma
The number N of equivalence classes of a set of colourings C in
the presence of a group of symmetries G is
1 X
N= |Cπ |
|G |
π∈G

where Cπ = {c ∈ C | π ∗ (c) = c} is the invariant set of π in the set

of colorings C .

If C is the set of all possible colourings with m colours and π is a

cyclic structure made of p cycles, then |Cπ | = mp .
For instance:
|C(1,2)(3,4) | = m2
|C(1)(2)(3)(4) | = m4
|C(1)(2,4)(3) | = m3
Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 2 / 25
Cycle index of a group

Assumption: G is a group of n-permutations, and π ∈ G

If π has type λ = [λ1 , λ2 , . . . , λn ] then
n
Y
Mπ = Mπ (x1 , x2 , . . . , xn ) = xiλi
i=1

where x1 , . . . , xn are unknowns.

The cycle index of G is
1 X
PG (x1 , x2 , . . . , xn ) = Mπ (x1 , x2 , . . . , xn ).
|G |
π∈G

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 3 / 25

Cycle index of a group
Example

The dihedral group G = D4 has 8 permutations, and:

M(1)(2)(3)(4) = x14 ,
M(1,3)(2)(4) = M(1)(2,4)(3) = x12 x2 ,
M(1,2)(3,4) = M(1,3)(2,4) = M(1,4)(2,3) = x22 ,
M(1,2,3,4) = M(1,4,3,2) = x4 .

If we add these terms and divide the sum by their number, we

obtain the cycle index of D4 :
1
PD4 (x1 , x2 , x3 , x4 ) = (x14 + 2 x12 x2 + 3 x22 + 2 x4 ),
8

Similarly, for the group C4 we obtain

1
PC4 (x1 , x2 , x3 , x4 ) = (x14 + x22 + 2 x4 ).
4

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 4 / 25

Burnside’s Lemma and the Cycle Index

According to Burnside, the number of colourings of n objects with

m colors, by taking into account the symmetries of group G , is
N = PG (m, m, . . . , m).
Example
The number of 4-beads necklaces with m colors is
1
PD4 (m, m, m, m) = (m4 + 2 m3 + 3 m2 + 2 m).
8
because we already know that
1
PD4 (x1 , x2 , x3 , x4 ) = (x14 + 2 x12 x2 + 3 x22 + 2 x4 )
8

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 5 / 25

Burnside’s Lemma
Application

Q: How many 20-beads necklaces can be made by using 3 colors?

A: We compute the cycle index of the symmetry group D20 . D20 has 20 rotations:
The rotation with 0◦ has type [20, 0, 0, . . . , 0] ⇒ monomial x120
8 rotations with k · 18◦ where k ∈ {1, 3, 7, 9, 11, 13, 17, 19} have type
[0, . . . , 0, 1] ⇒ monomial 8 x20
4 rotations with k · 18◦ where k ∈ {2, 6, 14, 18} have type [λ1 , . . . , λ20 ]
with λ10 = 2 and λj = 0 for all j 6= 10 ⇒ monomial 4 x10 2

◦
4 rotations with k · 18 where k ∈ {4, 8, 12, 16} have type [λ1 , . . . , λ20 ]
with λ5 = 4 and λj = 0 for all j 6= 5 ⇒ monomial 4 x54
2 rotations with k · 18◦ where k ∈ {5, 15} have type [λ1 , . . . , λ20 ] with
λ4 = 5 and λj = 0 for all j 6= 4 ⇒ monomial 2 x45
Rotation with 10 · 18◦ has type [0, 2, 0, . . .] ⇒ monomial x210
and 20 reflections
10 reflections around axes passing through midpoints of opposite edges of
the regular polygon have type [0, 10, 0, . . . , 0] ⇒ monomial 10 x210
10 reflections around axes passing through opposite nodes of the regular
polygon have type [λ1 , . . . , λ20 ] with λ1 = 2 and λ9 = 1 ⇒ 10 x12 x29

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 6 / 25

Burnside’s Lemma
Application

Q: How many 20-beads necklaces can be made by using 3 colors?

A: We compute the cycle index of the symmetry group D20 . D20 has 20 rotations:
The rotation with 0◦ has type [20, 0, 0, . . . , 0] ⇒ monomial x120
8 rotations with k · 18◦ where k ∈ {1, 3, 7, 9, 11, 13, 17, 19} have type
[0, . . . , 0, 1] ⇒ monomial 8 x20
4 rotations with k · 18◦ where k ∈ {2, 6, 14, 18} have type [λ1 , . . . , λ20 ]
with λ10 = 2 and λj = 0 for all j 6= 10 ⇒ monomial 4 x10 2

◦
4 rotations with k · 18 where k ∈ {4, 8, 12, 16} have type [λ1 , . . . , λ20 ]
with λ5 = 4 and λj = 0 for all j 6= 5 ⇒ monomial 4 x54
2 rotations with k · 18◦ where k ∈ {5, 15} have type [λ1 , . . . , λ20 ] with
λ4 = 5 and λj = 0 for all j 6= 4 ⇒ monomial 2 x45
Rotation with 10 · 18◦ has type [0, 2, 0, . . .] ⇒ monomial x210
and 20 reflections
10 reflections around axes passing through midpoints of opposite edges of
the regular polygon have type [0, 10, 0, . . . , 0] ⇒ monomial 10 x210
10 reflections around axes passing through opposite nodes of the regular
polygon have type [λ1 , . . . , λ20 ] with λ1 = 2 and λ9 = 1 ⇒ 10 x12 x29

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 6 / 25

Burnside’s Lemma
Application

Q: How many 20-beads necklaces can be made by using 3 colors?

A: We compute the cycle index of the symmetry group D20 . D20 has 20 rotations:
The rotation with 0◦ has type [20, 0, 0, . . . , 0] ⇒ monomial x120
8 rotations with k · 18◦ where k ∈ {1, 3, 7, 9, 11, 13, 17, 19} have type
[0, . . . , 0, 1] ⇒ monomial 8 x20
4 rotations with k · 18◦ where k ∈ {2, 6, 14, 18} have type [λ1 , . . . , λ20 ]
with λ10 = 2 and λj = 0 for all j 6= 10 ⇒ monomial 4 x10 2

◦
4 rotations with k · 18 where k ∈ {4, 8, 12, 16} have type [λ1 , . . . , λ20 ]
with λ5 = 4 and λj = 0 for all j 6= 5 ⇒ monomial 4 x54
2 rotations with k · 18◦ where k ∈ {5, 15} have type [λ1 , . . . , λ20 ] with
λ4 = 5 and λj = 0 for all j 6= 4 ⇒ monomial 2 x45
Rotation with 10 · 18◦ has type [0, 2, 0, . . .] ⇒ monomial x210
and 20 reflections
10 reflections around axes passing through midpoints of opposite edges of
the regular polygon have type [0, 10, 0, . . . , 0] ⇒ monomial 10 x210
10 reflections around axes passing through opposite nodes of the regular
polygon have type [λ1 , . . . , λ20 ] with λ1 = 2 and λ9 = 1 ⇒ 10 x12 x29
1
I PD20 (x1 , x2 , . . . , x20 ) = 40 (x120 + 10 x12 x29 + 11 x210 + 2 x45 + 4 x54 + 4 x10
2 + 8x )
20
⇒ N = P20 (3, . . . , 3) = 87 230 157
Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 6 / 25
Applications of the cycle index
Pólya’s enumeration formula

The cycle index can be used to solve more complicated problems to

count arrangements in the presence of symmetries. For instance:
How can we find the number of equivalence classes of colourings of
arrangements of n objects with m colours y1 , y2 , . . . , ym , if every
colour should appear a predefined number of times?

Definition (Pattern Inventory)

The pattern inventory of the colourings of n objects with m colours in
the presence of symmetries from a group G is the polynomial
X
FG (y1 , y2 , . . . , ym ) = av y1n1 y2n2 . . . ymnm
v

where
the sum is over all vectors v = (n1 , n2 , . . . , nm ) of positive integers
such that n1 + n2 + . . . + nm = n, and
a(n1 ,n2 ,...,nm ) is the number of non-equivalent colourings of these n
objects, where every colour yi appears exactly ni times.
Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 7 / 25
Pólya’s Enumeration Formula

Example
How many different necklaces can be made with 2 red beads (r ), 9 black
(b) and 9 white (w )? We assume that the symetries of this necklace are
the permutations of the dihedral group D20 , made of
I 20 rotations
I 20 symmetries

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 8 / 25

Pólya’s Enumeration Formula

Example
How many different necklaces can be made with 2 red beads (r ), 9 black
(b) and 9 white (w )? We assume that the symetries of this necklace are
the permutations of the dihedral group D20 , made of
I 20 rotations
I 20 symmetries
Answer: This is the coefficient of r 2 b 9 w 9 in the pattern inventory, which
is the polynomial
X X
FD20 (r , b, w ) = av r i b j w k = a(i,j,k) r i b j w k .
v=(i,j,k) i+j+k=20
i+j+k=20 i,j,k≥0
i,j,k≥0

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 8 / 25

Pólya’s Enumeration Formula

Example
How many different necklaces can be made with 2 red beads (r ), 9 black
(b) and 9 white (w )? We assume that the symetries of this necklace are
the permutations of the dihedral group D20 , made of
I 20 rotations
I 20 symmetries
Answer: This is the coefficient of r 2 b 9 w 9 in the pattern inventory, which
is the polynomial
X X
FD20 (r , b, w ) = av r i b j w k = a(i,j,k) r i b j w k .
v=(i,j,k) i+j+k=20
i+j+k=20 i,j,k≥0
i,j,k≥0

In 1937, G. Pólya found a simple formula to compute the pattern

inventory, using the cycle index of the group. (see next slide)

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 8 / 25

Pólya’s Enumeration Formula

Theorem
Suppose S is an arrangement of n objects colorable with m colors
y1 , . . . , ym , and G is a group of n-permutations. Let
1 X
PG (x1 , x2 , . . . , xn ) = Mπ (x1 , x2 , . . . , xn )
|G |
π∈G

be the cycle index of G . The pattern inventory of all colourings of the

objects of S with colours y1 , . . . , ym in the presence of symmetries of G is
m m m
!
X X X
FG (y1 , . . . , ym ) = PG yi , yi2 , . . . , yin .
i=1 i=1 i=1

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 9 / 25

Pólya’s Enumeration Formula
Applications

The pattern inventory of colourings FG (r , g , b) with red (r ) green (g )

and blue (b) of the beads of a necklace with 4 beads (=square vertices)
in the presence of symmetries from G = D4 can be computed as follows:
m = 3 because the set of colours is {r , g , b}
The P cycle index is PD4 (x1 , x2 , x3 , x4 ) =
1 1 4 2 2
|D4 | π∈D4 Mπ (x1 , x2 , x3 , x4 ) = 8 (x1 + 2 x1 x2 + 3 x2 + 2 x4 )

FG (r , g , b) =PD4 (r + g + b, r 2 + g 2 + b 2 , r 3 + g 3 + b 3 , r 4 + g 4 + b 4 )
1
= (r + g + b)4 + 2 (r + g + b)2 (r 2 + g 2 + b 2 )
8
+ 3 (r 2 + g 2 + b 2 )2 + 2 (r 4 + g 4 + b 4 )

=r 4 + g 4 + b 4 + r 3 g + r g 3 + r 3 b + r b 3 + g 3 b + g b 3
+ 2 r 2 g 2 + 2 r 2 b2 + 2 g 2 b2 + 2 r 2 g b + 2 r g 2 b + 2 r g b2

E.g., there are 2 colorings with 1 red bead, 1 green, and 2 blue.
Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 10 / 25
Pólya’s Enumeration Formula
Applications

The pattern inventory of colourings FG (r , g , b) with red (r ) green (g )

and blue (b) of the beads of a necklace with 4 beads (=square vertices)
in the presence of symmetries from G = C4 can be computed as follows:
m = 3 because the set of colourings is {r , g , b}
The cycle index is
1 X 1
PC4 (x1 , x2 , x3 , x4 ) = Mπ (x1 , x2 , x3 , x4 ) = (x14 + x22 + 2 x4 )
|C4 | 4
π∈C4

FG (r , g , b) =PC4 (r + g + b, r 2 + g 2 + b 2 , r 3 + g 3 + b 3 , r 4 + g 4 + b 4 )
1
= (r + g + b)4 + (r 2 + g 2 + b 2 )2 + 2 (r 4 + g 4 + b 4 )


4
=r 4 + g 4 + b 4 + r 3 g + r g 3 + r 3 b + r b 3 + g 3 b + g b 3
+ 2 r 2 g 2 + 2 r 2 b2 + 2 g 2 b2 + 3 r 2 g b + 3 r g 2 b + 3 r g b2

E.g., there are 3 colourings with 1 red bead, 1 green, and 2 blue.

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 11 / 25

Stirling cycle numbers

Problem
In how many ways can n persons be seated at k round tables, such that
no table is unoccupied? At every table can stay any number o persons
between 1 and n.

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 12 / 25

Stirling cycle numbers

Problem
In how many ways can n persons be seated at k round tables, such that
no table is unoccupied? At every table can stay any number o persons
between 1 and n.
Answer: Every answer to this problem is described by a cycle structure
with k disjoint structures C1 . . . Ck where Ci is the cycle describing the
people seated at table i.

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 12 / 25

Stirling cycle numbers

Problem
In how many ways can n persons be seated at k round tables, such that
no table is unoccupied? At every table can stay any number o persons
between 1 and n.
Answer: Every answer to this problem is described by a cycle structure
with k disjoint structures C1 . . . Ck where Ci is the cycle describing the
people seated at table i.
Example
The cycle structure (1, 2, 4)(3, 6, 9, 10)(5)(7, 8) represents a possible
arrangement of 10 persons at 4 round tables:
The people at one table are arranged 1,2,4 clockwise.
The people at another table are arranged 3,6,9,10 clockwise.
At another table stays only person 5.
At the remaining table are persons 7 and 8.

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 12 / 25

Stirling cycle numbers

Definition
The Stirling cycle number kn is the number of possibilities to seat
 

n persons at k identical round tables such that no round table is

left unoccupied.

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 13 / 25

Stirling cycle numbers

Definition
The Stirling cycle number kn is the number of possibilities to seat
 

n persons at k identical round tables such that no round table is

left unoccupied.

From the previous remark results that kn is the number of

 

n-permutations whose cycle structure is made of exactly k cycles.

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 13 / 25

Stirling cycle numbers

Definition
The Stirling cycle number kn is the number of possibilities to seat
 

n persons at k identical round tables such that no round table is

left unoccupied.

From the previous remark results that kn is the number of

 

n-permutations whose cycle structure is made of exactly k cycles.

Question: How to compute directly kn ?
 

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 13 / 25

Stirling cycle numbers

Definition
The Stirling cycle number kn is the number of possibilities to seat
 

n persons at k identical round tables such that no round table is

left unoccupied.

From the previous remark results that kn is the number of

 

n-permutations whose cycle structure is made of exactly k cycles.

Question: How to compute directly kn ?
 

Answer: Identify a recursive definition for Stirling cycle

numbers, and then solve it.

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 13 / 25

Stirling cycle numbers
Obvious properties

1. We can not place n persons at 0 tables, unless n = 0 (in this

special case, the number is assumed to be 1). Thus
  
n 1 if n = 0,
=
0 0 if n > 0.
2. n ≥ 1 persons can be seated at 1 table in (n − 1)! ways. Thus:
 
n
= (n − 1)! if n ≥ 1.
1
3. n persons can be seated at n tables
 in just 1 way: every
person is alone at a table. Thus: nn = 1.
4. n persons can be seated at n − 1 tables as follows: all persons,
except one couple, stay alone at a table. Thus
   
n n
= number of possible couples = .
n−1 2
Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 14 / 25
Stirling cycle numbers
Obvious properties

5. If the number of tables k is negative or if there are more

tables than persons, the problem has no solution. Thus:
 
n
= 0 if k < 0 or k > n.
k

6. Every permutation has a cycle structure made of k cycles,

where 1 ≤ k ≤ n. According to the rule of sum
n  
X n
= n!
k
k=1

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 15 / 25

Stirling cycle numbers
Finding a recurrence relation

How can we seat n > 0 persons at k > 0 round tables?

We distinguish two disjoint cases:
1 Place the first n − 1 persons at k − 1 round tables, and

afterwards place
n−1person
 n at table k. This case can be
performed in k−1 ways.
2 Place n − 1 persons at k round tables, and afterwards add

person n together with other persons at a round table.

Placing n − 1 persons at k tables can be done in n−1
 
k ways.
Placing person n at a round table = placing person n to the
left of one of the other persons i ∈ {1, 2, . . . , n − 1} ⇒ n − 1
ways.
⇒ This case can be performed in (n − 1) · n−1
 
k ways.
According to the rule of sum
     
n n−1 n−1
= (n − 1) + if n ≥ 1 and k ≥ 1.
k k k −1
Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 16 / 25
Stirling cycle numbers
Comparison with binomial numbers

We already P
know that

the binomial formula holds
(x + y )n = nk=0 kn x k y n−k . For y = 1 we get:
n  
n
X n
(x + 1) = xk
k
k=0

Also, in a previous lecture we gave a combinatorial proof that

     
n n−1 n−1
= + .
k k k −1
We just proved combinatorial proof that
     
n n−1 n−1
= (n − 1) + .
k k k −1

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 17 / 25

Stirling cycle numbers
Comparison with binomial numbers

We already P
know that

the binomial formula holds
(x + y )n = nk=0 kn x k y n−k . For y = 1 we get:
n  
n
X n
(x + 1) = xk
k
k=0

Also, in a previous lecture we gave a combinatorial proof that

     
n n−1 n−1
= + .
k k k −1
We just proved combinatorial proof that
     
n n−1 n−1
= (n − 1) + .
k k k −1
We want to get a formula for Stirling cycle numbers, which is
similar to the binomial formula.
Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 17 / 25
Stirling cycle numbers
Identifying a generative function
X n 
x k . Then G0 (x) = 00 x 0 = 1 · 1 = 1, and for

Let Gn (x) =
k
k
n≥1
X n 
Gn (x) = xk
k
k
X n − 1 X n − 1
k
=(n − 1) x + xk
k k −1
k k
=(n − 1)Gn−1 (x) + x Gn−1 (x)
=(x + n − 1)Gn−1 (x)

⇒ Gn (x) = x · (x + 1) · (x + 2) · . . . · (x + n − 1) .
| {z }
  notation: x n̄
X n
Thus x n̄ = xk.
k
k
Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 18 / 25
Stirling cycle numbers
The triangle of Stirling cycle numbers

This is an infinite triangle of Stirling cycle numbers growing downwards:

n 
k k=0 1 2 3 4 5 6 7 8 n!
n=0 1 1
1 0 1 1
2 0 1 1 2
3 0 2 3 1 6
4 0 6 11 6 1 24
5 0 24 50 35 10 1 120
6 0 120 274 225 85 15 1 720
7 0 720 1764 1624 735 175 21 1 5040
8 0 5040 13068 13132 6769 1960 322 28 1 40320

Recursive formula used in the computation:

     
n n−1 n−1
= (n − 1) + .
k k k −1

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 19 / 25

Binomial numbers
The triangle of binomial numbers

This is an infinite triangle of binomial numbers growing downwards:

n


k k=0 1 2 3 4 5 6 7 8 n!
n=0 1 1
1 1 1 1
2 1 2 1 2
3 1 3 3 1 6
4 1 4 6 4 1 24
5 1 5 10 10 5 1 120
6 1 6 15 20 15 6 1 720
7 1 7 21 35 35 21 7 1 5040
8 1 8 28 56 70 56 28 8 1 40320

Recursive formula used in the computation:

     
n n−1 n−1
= + .
k k k −1

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 20 / 25

Stirling set numbers

Problem
In how many ways can we divide n persons in k non-empty and disjoint
groups, if the order of persons in one group does not matter?

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 21 / 25

Stirling set numbers

Problem
In how many ways can we divide n persons in k non-empty and disjoint
groups, if the order of persons in one group does not matter?

Example
The set {1, 2, 3} can be partitioned in 2 non-empty subsets in 3 ways:
{1, 2}, {3}; {1, 3}, {2}; and {1}, {2, 3}.

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 21 / 25

Stirling set numbers

Problem
In how many ways can we divide n persons in k non-empty and disjoint
groups, if the order of persons in one group does not matter?

Example
The set {1, 2, 3} can be partitioned in 2 non-empty subsets in 3 ways:
{1, 2}, {3}; {1, 3}, {2}; and {1}, {2, 3}.

Definition
The number of ways in which we can partition a set of n elements in
exactly k non-empty and disjoint subsets is the Stirling set number kn .
Often in the literature this number is denoted by S(n, k) instead of kn .

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 21 / 25

Stirling set numbers
Obvious properties

1. There is only one way to place n people in one group, and also only
one way to split n people in n groups. Thus:
   
n n
= = 1.
1 n
2. We can not place n > 0 people in 0 groups. If n = 0 then we
assume there there is 1 way to place 0 people in 0 groups. Thus:
  
n 1 if n = 0,
=
0 0 if n > 0.
3. Splitting n people in n − 1 groups amounts to choosing a couple of
persons for one group; all other persons are alone in their group.
Thus    
n n
= .
n−1 2
4. It is obvious that
 
n
=0 if k < 0 or k > n.
k
Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 22 / 25
Stirling set numbers
Finding a recurrence relation

How can we split n > 0 persons in k > 0 non-empty and disjoint

subsets?
We distinguish 2 disjoint cases:
1. We split the first n − 1 persons in k − 1 groups;
n−1 then person n
is obliged to form a singleton group {n} ⇒ k−1 possibilities.
n−1
2. We split the first n − 1 persons in k groups ⇒ k
possibilities; afterwards, we add person n to one of those k
groups⇒ k · n−1

k possibilities.
According to the rule of sum
     
n n−1 n−1
=k· + if n ≥ 1 and k ≥ 1.
k k k −1

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 23 / 25

Stirling set numbers
The triangle of Stirling set numbers

This is an infinite triangle of Stirling set numbers growing downwards:

n
k k=0 1 2 3 4 5 6 7 8
n=0 1
1 0 1
2 0 1 1
3 0 1 3 1
4 0 1 7 6 1
5 0 1 15 25 10 1
6 0 1 31 90 65 15 1
7 0 1 63 301 350 140 21 1
8 0 1 127 966 1701 1050 266 28 1

Recursive formula used in the computation:

     
n n−1 n−1
=k· + .
k k k −1

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 24 / 25

References

1 J. M. Harris, J. L. Hirst, M. J. Mossinghoff. Combinatorics

and Graph Theory, Second Edition. Springer 2008.
§2.7. Pólya’s Theory of Counting.
2 G. Pólya. Kombinatorische Anzahlbestimmungen für Gruppen,
Graphen, und chemische Verbindungen, Acta Math. 68
(1937), 145–254; English transl. in G. Pólya and R. C. Read,
Combinatorial Enumeration of Groups, Graphs, and Chemical
Compounds (1987).

Isabela Drămnesc UVT Graph Theory and Combinatorics – Lecture 6 25 / 25