You are on page 1of 3

Tutorial 1 Solutions

MATH 239 Winter 2024

Week of January 15

T-1-1. By counting the same set in two different ways, establish combinatorially the identity
   
n n−1
7 =n .
7 6

Solution. Let’s rewrite the RHS as n1 n−1


 
6 . This is the number of ways of choosing 1 element
from an n-element set, and then a 6-element subset from the remaining n − 1 elements. In
other words, it counts the number of elements of the form (a, U ) where a ∈ {1, . . . , n} and
U ⊆ {1, . . . , n} \ {a}.
Another way of thinking of this is that we have a 7-element subset of an n-element set,
and we’ve marked one of those 7 elements as “special”. There are n7 ways of choosing


a 7-element subset of an n-element, set, and then 71 = 7 ways of choosing one of those 7


elements to be special. □

T-1-2. For n ≥ 1, give a bijective proof that


X n X  n
=
k k
k is even k is odd

Solution. Let’s start by consider an example of a simple case: n = 3, with the set S = {1, 2, 3}.
We can see that the LHS is counting all the even-sized subsets of S, which are: {}, {1, 2}, {1, 3}, {2, 3}.
And the RHS is counting all the odd-sized subsets of S, which are {1}, {2}, {3}, {1, 2, 3}.
Can we find a bijection between these? One such bijection is the set complement: {} maps to
{1, 2, 3}, {1, 2} maps to {3}, and so on. We could easily extend this to work for general odd n.
But it won’t work for even n: if we have an even-sized subset of {1, . . . , n} for an even n, the
complement of that even-sized subset will also be even. So we have to try something else for
our general solution.
Let’s try an example with n = 2. The LHS counts the even sized subsets {}, {1, 2}, and the
RHS counts the odd-sized subsets {1}, {2}. Can you think of a general mapping between
them? The mapping we’ll use is: if 1 is in the subset, remove it; if 1 is not in the subset, add it.
Clearly this turns an odd-sized subset into an even-sized subset and vice versa. Moreover,
it’s clearly invertible.

1
MATH 239 Winter 2024 Tutorial 1 Solutions

To write it mathematically, let


(
u ∪ {1}, if 1 ̸∈ u
f (u) =
u \ {1}, if 1 ∈ u

If u is an even-sized subset, then f (u) is an odd-sized subset, and vice versa. Moreover,
f −1 (f (u)) = u and f (f −1 (u)) = u for all u ⊆ {1, . . . , n} for all n. □

T-1-3. By counting the same set in two different ways, establish combinatorially the identity
n X
k−1  
X k−1
= 2n − 1.
j
k=1 j=0

Solution. Let’s dissect the LHS.


We’ve seen k−1

j before: it’s the number of j-element subsets of a k − 1-element set.
Pk−1
So j=0 is the number of subsets (of any size) of a k − 1-element set.
(We’d be tempted to recognize this sum as 2k−1 , and then note that nk=1 2k−1 = n−1 k
P P
k=0 2 =
2n − 1. But we’re asked for a combinatorial proof, so we should avoid such algebraic
manipulations and see if we can get further with counting sets.)
Looking at the RHS, we see 2n − 1, which is close to something we’ve seen before. We know
that 2n is the number of subsets of an n-element set. So 2n − 1 is the number of subsets of an
n-element set, excluding one of them. But which one? Well, the two obvious choices would
be either the empty set or the complete set. Noticing that the sum on the LHS goes from k = 1
to n, leaving out k = 0, perhaps that suggests we should think of the RHS as omitting the
empty set.
Coming back to the LHS, we have established that k−1
P
j=0 is the number of subsets (of any
size) of a k − 1-element set, and then we’re summing over k = 1 to n. We want to show
how this counts the number of non-empty subsets of an n-element set. One way to see this
is, for each k = 1 to n, construct a subset of {1, . . . , n} by including k, and then taking any
subset of {1, . . . , k − 1}. The number of subsets constructed in this way is the LHS. Any
subset constructed in this way will be non-empty. Moreover, all subsets of {1, . . . , n} can be
constructed in this way. So we have counted the same set as we were counting on the RHS,
and thus LHS=RHS.

Extra Practice.

EP-1-1. For n, t ∈ N with t ≥ 1, we can write the set S of multisets of size n with t types as
t
( )
X
S = (m1 , . . . , mt ) : m1 , . . . , mt ∈ N, mi = n .
i=1

2
MATH 239 Winter 2024 Tutorial 1 Solutions

(a) Let k ≥ 1 be an integer. Let T be the set of multisets of size n with t types such that for
all i ∈ {1, . . . , t}, the number of elements of i-th type is a multiple of k. Write T in set
notation. What is |T |?
(b) Let k ≥ 0 be an integer. Let U be the set of multisets of size n with t types such that for
all i ∈ {1, . . . , t}, the number of elements of i-th type is at least k. Write U in set notation.
What is |U |?

In each case, justify your answer using a combinatorial proof (for example, by constructing a
bijection to a suitable set).
Solution.

(a)
t
( )
X
T = (km1 , . . . , kmt ) : m1 , . . . , mt ∈ N, kmi = n
i=1

and |T | = 0 if k does not divide n; otherwise, |T | = n/k+t−1



t−1 . We can see this by

constructing a bijection f from T to the set T of multisets of size n/k with t types, defined
as f ((km1 , . . . , kmt )) = (m1 , . . . , mt ). Its inverse is f −1 ((m1 , . . . , mt )) = (km1 , . . . , kmt ).
(Be sure to check that these are mutually inverse bijections.)
(b)
t
( )
X
U= (m1 , . . . , mt ) : m1 , . . . , mt ∈ N≥k , mi = n ,
i=1

and |U | = 0 if kt > n; otherwise |U | = n−kt+t−1



t−1 . To see this, we construct a bijection f

from U to the set U of multisets of size n − kt with t types, defined as f ((m1 , . . . , mt )) =
(m1 − k, . . . , mt − k). Its inverse is f −1 ((n1 , . . . , nt )) = (n1 + k, . . . , nt + k). (Be sure to
check that these are mutually inverse bijections.)

You might also like