You are on page 1of 8

Binomial Coefficients

CSE 215 - Foundations in Computer Science

Dennis Wong
Binomial Coefficients
A binomial expression is the sum of two terms x + y.

We call the expression C(n, r) a binomial coefficient such values


n
occur as the coefficients in the expansion (x + y) .

Example: Consider the expansion of (x + y)3.

(x + y)3 = 1 x3 + 3 x2y + 3 xy2 + 1 y3



3 2 2 3

= C(3, 0) x + C(3, 1) x y + C(3, 2) xy + C(3, 3) y

We can think of the coefficient C(3, 1) as the number of ways to


choose one y from each of the three terms (x + y)(x + y)(x + y).

Binomial Theorem
n n n-j j

Theorem: (x + y) = ∑ C(n, j) x y
j=0
n n-1 n-1 n
= C(n, 0) x + C(n, 1) x y + … + C(n, n-1) xy + C(n, n)y 


12 13 25
Example 1: What is the coefficient of x y in the expansion of (x + y) ?


Solution: C(25, 13)

12 13 25
Example 2: What is the coefficient of x y in the expansion of (2x - 3y) ?

25 25 25 25-j j

Solution: (2x - 3y) = (2x + (-3y)) = ∑ C(n, j) (2x) (-3y)


j=0
25 25-j 25-j j j

= ∑ C(n, j) 2 x (-3) y
j=0
12 13
Thus the coefficient of x y when j = 13:

12 13
C(25, 13) 2 (-3)
Binomial Theorem
n
Corollary 1: ∑ C(n, j) = 2n

j=0

Proof: We apply the binomial theorem by setting x and y


equal to 1. 


Combinatorial proof of Corollary 1: There are 2n binary
strings with length n. Partition these strings into groups
depending on the number of 1’s in the string: 0 ones, 1 one,
2 ones, 3 ones, …, n ones. The number of length n strings
with exactly j ones is C(n, j). Thus:

n
∑ C(n, j) = 2n

j=0
Binomial Theorem
n
Corollary 2: ∑ (-1)j C(n, j) = 0

j=0

Proof: We apply the binomial theorem by setting x = 1 and y = -1. 




Observe that this means that:

C(n, 0) + C(n, 2) + … = C(n, 1) + C(n, 3) + …

This also means that the number of subsets with an odd number of
elements equals the number of subsets with an even number of
elements.

Does the number of length n bitstrings with an even number of ones


equal the number of strings with an odd number of ones?

Pascal's Identity

Pascal’s Identity: C(n + 1, k) = C(n, k - 1) + C(n, k)

Combinatorial Proof: The expression C(n + 1, k) counts the number of


binary strings of length n + 1 with k ones. We can partition these strings
into 2 groups:

(1): strings that begin with 1



(2): strings that begin with 0

In (1), since the strings start with 1, that means the remainder of each
string has k - 1 ones: there are C(n, k - 1) such strings of length n.

In (2), since the strings start with 0, that means the remainder of each
string has k ones: there are C(n, k) such strings of length n.

Therefore: C(n + 1, k) = C(n, k - 1) + C(n, k).


Vandermonde’s Identity
r
Pascal Identity: C(m + n, r) = ∑ C(m, r - k) C(n, k)

k=0

Combinatorial Proof: The expression C(m + n, r) counts to select


a total of r objects from two sets A and B, one of size m and one
of size n.

One way to pick r elements from these sets is to pick r - k


elements from set A, and k elements from set B.

There are C(m, r - k) C(n, k) ways to do this.

Finally, we sum over all possible values of k to obtain the identity.


One More Theorem
n
Theorem: C(n + 1, r + 1) = ∑ C(j, r)

j=r

Combinatorial Proof: The expression C(n + 1, r + 1) counts the


number of bit strings of length n + 1 with r + 1 ones.

Now we consider all the possible positions for the last one in
these strings.

The position can occur in position j where r + 1 ≤ j ≤ n + 1. The


number of strings with the last one occurs in position j is C(j - 1, r).

Summing over all possible values of j, we get:

n+1 n
∑ C(j - 1, r) = ∑ C(j, r)
j = r+1 j=r

You might also like