MATH2222 Homework 3

Isaac Leong
March 2023

Collaboration Statement
For Question 2, I discussed with Taij Mundy and Thomas Bridgart in the tutorial. All the
other questions are done individually.

1
Question 1
Prove : If it is a product of odd numbers, then the product is also an odd number.
Prove by induction on n, the number of odd numbers in the product.

Base case, n = 1,
Since a1 is an odd number, then a1 is an odd number. Therefore the base case is true.

Inductive step: Assume a1 a2 ...an is an odd number.

Prove: a1 a2 ...an an+1 is an odd number
By induction hypothesis, a1 a2 ...an is an odd number. Also, an+1 is an odd number. Be-
cause a product of two odd numbers is an odd number as (2m + 1)(2n + 1) ≡ 1 (mod 2),
therefore, (a1 a2 ...an )an+1 is also an odd number. Hence, a1 a2 ...an an+1 is an odd number.
This shows that if P (n) is true, then P (n + 1) is also true.

Therefore, by induction, product of odd numbers is also an odd number. ■

2
Question 2
Find and prove by induction a formula for the sum of the first k odd positive integers.
k
X
First, through guessing, the formula is 2i + 1 = (k + 1)2 .
i=0
Prove by induction.

Base case: when k = 0, then 2 × 0 + 1 = 1 = (0 + 1)2 . Therefore, the base case is true.
n
X Xn+1
Inductive step: Assume 2i + 1 = (n + 1)2 . Prove: 2i + 1 = (n + 2)2 .
i=0 i=0

n+1
X Xn
2i + 1 = ( 2i + 1) + 2n + 3
i=0 i=0
= (n + 1)2 + 2n + 3 (Induction Hypothesis)
= n2 + 2n + 1 + 2n + 3
= n2 + 4n + 4
= (n + 2)2

Therefore, if P (n) is true, then P (n + 1) is also true.

k
X
Therefore, by mathematical induction, 2i + 1 = (k + 1)2 is true for all k ∈ N. ■
i=0

3
Question 3
n
X
Let fn denotes the n-th Fibonacci numbers. Prove that: fi2 = fn fn+1
i=1
Prove by induction.
1
X
Base case: when n = 1, then fi2 = 1 = 1 · 1 = f1 f2 . Therefore, the base case is true.
i=1

n
X n+1
X
Inductive step: Assume fi2 = fn fn+1 is true. Prove: fi2 = fn+1 fn+2 .
i=1 i=1

n+1
X n
X
fi2 = fi2 + fn+1
2

i=1 i=1
2
= fn fn+1 + fn+1 (Inductive Hypothesis)
= fn+1 (fn + fn+1 )
= fn+1 fn+2 (Definition of Fibonacci numbers)

Therefore, this shows that if P (n) is true, then P (n + 1) is also true.

n
X
Therefore, by induction, fi2 = fn fn+1 is true for all n ∈ N. ■
i=1

4
Question 4
Prove: Tn < 2n for all n ≥ k + 1, where Tn is the ”k-bonacci numbers”.
Prove by strong induction on n.

Base case: Prove: Tk+1 < 2k+1 .

First, by Bernoulli inequality, 2k+1 = (1 + 1)k+1 ≥ 1 + k + 1 > k.
Also, since Tk+1 = k, therefore, 2k+1 > k = Tk+1 .
Therefore, the base case is true.

Inductive Step: Assume for all k < n + 1, Tk < 2k . Prove: Tn+1 < 2n+1 .

Tn+1 = Tn + Tn−1 + ... + Tn+1−k (Definition of ”k-bonacci numbers”)

n n−1 n+1−k
<2 +2 + ... + 2 (Induction Hypothesis)
n n−1
≤2 +2 + ... + 20
1 − 2n+1
= (Geometric series)
1−2
n+1
=2 −1
n+1
<2

Therefore, P (n + 1) is true whenever P (k), k < n + 1, is true.

Therefore, by strong induction, Tn < 2n for all n ≥ k + 1 is true. ■

5
Question 5
Suppose we draw n lines in the plane, chopping the plane into rn regions. Prove that
1
rn ≤ (n2 + n + 2)
2
.
Prove by induction on n.

Base case: when n = 0, meaning that there is no line in the plane. Then rn ≤ 12 (0+0+2) =
1, which is true because if there is no line, then the whole plane is 1 regions. Therefore,
there is one region.

Inductive step: Assume the inequality holds for n lines on the plane. Prove: The maxi-
mum number of regions after adding another line is 12 ((n + 1)2 + (n + 1) + 2).
The maximum number of regions can be created by adding an extra line to the original
configurations of lines when the newly added line intersects with all the other lines that
were already there. This implies that the number of newly created regions is n + 1 as there
were n lines originally.

Hence, rn+1 ≤ rn + n + 1 (From discussion above)

1
≤ (n2 + n + 2) + n + 1 (Induction hypothesis)
2
1 2
= (n + n + 2 + 2n + 2)
2
1
= ((n + 1)2 + (n + 1) + 2)
2

Therefore, rn+1 ≤ 12 ((n + 1)2 + (n + 1) + 2) is true whenever rn ≤ 12 (n2 + n + 2).

Therefore, by induction, rn ≤ 21 (n2 + n + 2). ■

6
Question 6
(i)
Prove: R can be tiled with L-tiles when R is a 2k by 2k chessboard with one corner 1 by 1
square removed. (k ≥ 1)
Prove by induction on k.

Base case(k = 1): This case is obviously true because it is just a single L-tile.

Inductive step: Assume that it is possible for R which is a 2k by 2k chessboard with one
corner 1 by 1 square removed. Prove: R′ which is a 2k+1 by 2k+1 chessboard with one cor-
ner 1 by 1 square removed can also have a tiling of L.
In order to show that the new board R′ does have a L-tiling, first divide the square into
4 equal sized sub-squares of size 2k by 2k (of course with one that has a corner missing).
Then, the upper right sub-square can be completely covered with L-tiles by induction hy-
pothesis. Then, place a L-tile around the bottom left corner of the upper right sub-square.
This results in the effect of taking a corner away from the remaining three 2k by 2k sub-
squares. Then, by the induction hypothesis again, these three remaining sub-squares also
has an L-tiling. This process is shown in the following diagram.

Figure 1: This is how the big square with size 2k+1 by 2k+1 is separated.

Therefore, as long as a corner-missing square of size 2k by 2k has L-tiling, a corner-missing

square of size 2k+1 by 2k+1 has L-tiling as well.

Therefore, by induction, it is true that R can be tiled with L-tiles when R is a 2k by 2k

chessboard with one corner 1 by 1 square removed. (k ≥ 1) ■

7
(ii)
Prove: R can be tiled with L-tiles when R is a 2k by 2k chessboard with a single 1 by 1
square removed. (k ≥ 1)
Prove by induction on k.

Base case(k = 1): This case is also true because any 2 by 2 square with any one 1 by 1
square removed is always in a L-shape configurations.

Inductive step: Assume that it is possible for R which is a 2k by 2k chessboard with one 1
by 1 square removed anywhere on the board. Prove: R′ which is a 2k+1 by 2k+1 chessboard
with one 1 by 1 square removed also has an L-tiling.
First, similar approach can be taken as in (i). We still separate the square into 4 equal
sub-squares with each of them has a side length of 2k . Then, since all these four sub-squares
are equivalent by simply rotating 90o , 180o and 270o , then assume that the empty 1 by 1
square is located somewhere in the top right sub-square. Then, by induction hypothesis,
this square with side length of 2k and a hole of 1 by 1 square has L-tiling. The remain-
ing three squares can be taken care in a similar approach as what is done in (i) as having
a corner missing is only a special case of any one 1 by 1 square missing. This shows that
the remaining three sub-squares in the big square also have L-tiling. Therefore, the whole
square of side length of 2k+1 with a single 1 by 1 square missing also has L-tiling.
Hence, it is shown that as long as a square of size 2k by 2k with a single 1 by 1 square
taken off has L-tiling, it is true that the square of size 2k+1 by 2k+1 with a single 1 by 1
square take off also has L-tiling.

Therefore, by induction, it is true that R can be tiled with L-tiles whenever R is a 2k by

2k chessboard with any single 1 by 1 square removed. ■

8
Problems for pleasure. Question 1
Recall the Fibonacci numbers {fn }n≥1 . Prove that for all n, m ≥ 1.

fn+m+1 = fm fn + fm+1 fn+1

.
Prove by induction on m.

Base case 1 (m = 1):

f1+n+1 = fn+2 = fn + fn+1

= 1 · fn + 1 · fn+1
= f1 fn + f2 fn+1

Therefore, the base case 1 is true.

Base case 2 (m = 2):

f2+n+1 = fn+3 = fn+2 + fn+1

= 1 · fn + 2 · fn+1
= f2 fn + f3 fn+1

Therefore, the base case 2 is true.

Inductive step: Assume the statement is true for m = k and m = k − 1.

fn+k+1 = fk fn + fk+1 fn+1

fn+k = fk−1 fn + fk fn+1

Prove: fn+k+2 = fk+1 fn + fk+2 fn+1 where m = k + 1.

fn+k+2 = fn+k+1 + fn+k (Definition of Fibonacci Numbers)

= fk fn + fk+1 fn+1 + fk−1 fn + fk fn+1
(Induction Hypothesis of m = k and m = k − 1)
= fn (fk−1 + fk ) + fn+1 (fk + fk+1 )
= fn fk+1 + fn+1 fk+2 (Definition of Fibonacci Numbers)

Therefore, fn+m+1 = fm fn + fm+1 fn+1 holds at m = k + 1 whenever the equation holds for
m = k and m = k − 1.

Therefore, the equation holds for all m ≥ 1 by principle of reduction. Also, notice that
there is no restriction placed on n throughout the proof. Therefore, this equation also
holds for all n ≥ 1.

Hence, this equation holds for all m ≥ 1 and n ≥ 1. ■