Combinatorics Note

Download as pdf or txt
Download as pdf or txt
You are on page 1of 2

Combinatorics Diary

Arya Vadnere
August 25, 2019

• Distribution Problem - n objects, k boxes;

Boxes Identical Boxes Distinct


 
k
Objects identical (Elbow notation)
n  
k
Objects identical; Each box at least 1
  n−k
Pn−1 k
Objects distinct i=0 S(n, k − i) kn
i    
n k n k
Onjects distinct; Each box at least 1 S(n, k) k − (k − 1) + (k − 2)n ...
1 2
• Functions Problem - No. of functions with domain an n-set and range an m-set -
No restriction One-one
No restriction mn m
Pn
n
Onto Pm × mm−n n!
• Combinations with
 repetitions
 allowed
 - x1 + x2 + ... + xn = k; xi ≥ 0.
n n−k+1
No. of ways = = . Method - Stars and separators.
k k
• We have      
n 0 n 1 n r
(1 − x)−n = x + x + ... + x + ...
0 1 r

• Flag Pole problem - P distinct poles, N distinct flags;


Order important Order important, each pole at least k Order unimportant Order unimportant, each pole at least k
 
(P −1+N )! P
(P −1)! × F! PN N!
(k!)P
× P N −P k
F −P ·k
 
n
• Theorem: Given n and k with n < k, and 0 ≤ m ≤ − 1, there exists a sequence of integers n > a1 ≥ a2 ≥ ... ≥
k
ak ≥ 0 s.t.      
a1 a2 a
m= + + ... + k
k k−1 1
(Proof - Bijection, lex ordering)

1
• Necklace Problem - Necklaces to be formed, with n beads, available in m colours. Two necklaces are identical if one
can be obtained from other by rotation. Then no. of distinct necklaces is
1 X
Nn = φ(r) mn/r
n
r|n

(Proof - Mobius Inversion, Blocks)


• Some Pigeonhole Theorems:
– Erdos-Szekeres Theorem - Let S = (a1 , a2 , ..., amn+1 ) be a sequence of mn + 1 real numbers. Then S has a
monotonically increasing subsequence with m + 1 terms or a strictly decreasing subsequence of n + 1 terms.

– Let n ∈ N, and let Nn = 1, 2, ..., 2n−1 . Let f : Nn → Nn s.t. f (1) = 1, and f (i) ≤ i ∀ i ∈ Nn . Then there
exists a sequence of elements
a1 < a2 < ... < an
satisfying
f (a1 ) ≤ f (a2 ) ≤ ... ≤ f (an )
 
p+q−2
– A good bound for Ramsey Numbers: R(p, q) ≤
p−1
• Stuff on probability
P (A∩B)
– Conditional probability - PA (B) = P (B) .

– Bayes’ Theorem - Let {A1 , A2 , ..., An } be a partition of the sample space Ω. Then

PA (B) P (Ak )
PB (Ak ) = Pn k
j=1 PAj (B) P (Aj )

– Poisson’s Theorem - Take an urn with b black balls and r white balls. Balls drawn successively and thrown out
of urn. Let Bn be probability that nth draw is black. Then
b
P (Bn ) =
b+r

– Random Variable - Any function X : Ω → R.


P
– Expected value of a random variable - E(X) = ω∈Ω X(ω) P (ω).
– (Linearity of Expectation) E (X1 + X2 + ... + Xn ) = E (X1 ) + E (X2 ) + ... + E (Xn ).
1 1
– Proposition - A randomly chosen permutation of {1, 2, ..., n} is expected to have 1 + 2 + 3 + ... + n1 .
• Solution of the Nim game -
– Consider a game as a digraph, with vertices as all possible positions, and edges leading from one position to
another. Kernel of a game is defined as a set of vertices s.t. there is no edge connecting two vertices in the set,
and from every vertex outside the set, there is an edge connecting it to a vertex inside the set.
Pn
– Define an operation bitwise XOR (⊕) on integers. Define ⊕ i=1 ai = a1 ⊕ a2 ⊕ ... ⊕ an .
– Notation - Let the Nim game have n piles, and ith pile having ai elements. Each position is thus represented as
{a1 , a2 , ..., an }.
Pn
– Nim Theorem - The Kernel of the Nim game is the set K = {{x1 , x2 , ..., xn } | ⊕ i=1 ai = 0}.

You might also like