You are on page 1of 3

CIS 160

Recitation Guide - Week 4


Topics Covered: Induction, Combinatorial Proofs

Problem 1: Prove using induction that for any positive integer n and for any d0 , d1 , . . . , dn−1 ∈
[0..9] we have:
n−1
X
dj · 10j < 10n
j=0

Solution:
Base Case: n = 1.
0
X
dj · 10j = dj · 100 = dj < 10 = 101
j=0

dj can only take values from 0 through 9, thus dj < 10. This concludes the Base Case.
Induction Hypothesis: Assume that the claim is true when n = k, for some integer k ≥ 1 such
that:
k−1
X
dj · 10j < 10k
j=0

Induction Step: We now want to show that our claim holds when n = k + 1. In other words, we
seek to show that:
Xk
dj · 10j < 10k+1
j=0

We see that we can show this as follows:


k
X k−1
X
dj · 10j = dk · 10k + dj · 10j (splitting the sum)
j=0 j=0

< dk · 10k + 10k (by IH)


k
= (dk + 1) · 10
≤ (9 + 1) · 10k
= 10k+1

Thus, we have shown our claim is true when n = k + 1, concluding our Induction Step and
completing our proof.

1
Problem 2: A set S is natural if S ⊂ N. Prove that if the maximum element in a natural set B
is n, then the power set of B has at most 2n+1 elements.
Solution:
Let P (k) be true if and only if the claim holds for natural sets with maximum element k.
Base Case: When k = 0, B = {0}. P(B) = {∅, {0}} and |P(B)| = 2. 20+1 = 2, 2 ≤ 2. Hence,
P (0) is true.
Induction Hypothesis: Assume P (k) is true for some k ≥ 0, k ∈ N.
Bogus Induction Step: To prove that P (k + 1) is true, consider a natural set with maximum
element k. The cardinality of its power set is at most 2k+1 by the induction hypothesis. Add k + 1
to the set. The cardinality of the power set of our new natural set is at most 2k+1 ∗ 2 = 2k+2
because for each element of our old power set, we can either include or not include k + 1. This
doubles the cardinality. Thus, P (k + 1) is true.
Why is this bogus? If we start with a natural set whose maximum element is k, we know that the
set contains k. When we added k + 1 to the set, our maximum element became k + 1 but our set
by definition still contains k. Having a natural set with k + 1 as the maximum element does not
necessarily mean that k is also in the set and thus the proof is not correct, since we haven’t covered
cases where the maximum element is k + 1 but k isn’t in the set. For example, we haven’t covered
the following case: {0, 2, k + 1}, where k > 3.
Induction Step: To prove that P (k + 1) is true, consider a natural set A with maximum element
k + 1. Let us remove the maximum element k + 1 of this set to construct a new set B, so that
B = A \ {k + 1}. We now have two cases:
Case 1: k ∈ B
If k ∈ B then it must be the maximum element in B. Therefore we can apply our induction
hypothesis and conclude that |P(B)| ≤ 2k+1 . We can now construct P(A) from P(B). Note that
for each subset x ∈ P(B), x ∈ P(A) and x∪{k +1} ∈ P(A) – this means that each element in P(B)
have exactly 2 corresponding elements in P(A) (one without k + 1 and one with it). Hence

|P(A)| = 2 · |P(B)| ≤ 2 · 2k+1 = 2k+2

Case 2: k 6∈ B
If k 6∈ B then it must have have another maximum element ` < k. Let us construct a new set
C = B ∪ {k}. Note that C must have k as its maximum element by construction. Applying the
induction hypothesis, we know that |P(C)| ≤ 2k+1 . Now let us consider D = C ∪ {k + 1}. By
similar arguments to that in Case 1, we have that

|P(D)| = 2 · |P(C)| ≤ 2 · 2k+1 = 2k+2

Lastly, note that A = D \ {k}, so we necessarily have that |A| ≤ |D| (we can actually be stricter
here, but we don’t care at this point since we have what we want.) Therefore we have:

|P(A)| ≤ |P(D)| ≤ 2k+2

We have shown P(A) ≤ 2k+2 in all cases and thus proven the induction step.

2
Problem 3:
Give a combinatorial proof for the following, where m ≤ n:
m     
X n n−k n
= 2m
k m−k m
k=0

Solution:
Consider the following counting problem:
Given a set of n people, how many ways are there to hire m people such that any
number of those m people can also be designated as managers?
RHS: We use two steps to hire people and designate managers as follows.
Step 1: Choose the m people who are hired out of the n job applicants.
Step 2: Designate any number of the m people as managers.
n

In Step 1, we are simply choosing m out of n items. Thus there are m ways to do Step 1. In Step
2, we are taking a subset m
n
 of m items, so there are 2 ways to do Step 2. Applying the multiplication
rule, there are 2m m ways to hire people and designate managers, which is the RHS.
LHS: Let S be a set that includes all of the ways that we can hire people and designate managers.
We can partition S into sets S0 , S1 , S2 , . . . , Sm where set Sk (0 ≤ k ≤ m) represents all of the ways
that we can hire m people and designate exactly k managers. For each k, |Sk | can be calculated as
follows.
Step 1: Choose the k managers that we want from the n total people.
Step 2: Hire people from the remaining n − k people so that we end up with a total of
m employees.
In Step 1, we are simply choosing k people out of n, so there are nk ways to do Step 1. In Step


2, we must hire additional people so that we have a total of m employees. Since we have already
hired k people, we can only hire m − k more people. In addition, we cannot choose any of those
 they have already been hired), so there are n − k people to choose from.
k people to hire (since
n−k
Thus, there are m−k ways to do Step 2.

Applying the multiplication rule, there are nk m−k


 n−k 
ways to hire m people and designate exactly
k managers.
Thus, the total number of ways to hire m people and designate any number of managers is
m m   
X X n n−k
|S| = |Sk | =
k m−k
k=0 k=0

which is the LHS.

You might also like