Tobias Rossmann
tobias.rossmann@universityofgalway.ie
Let's prove that
\[\frac 1 2+ \frac 1 4 + \dotsb + \frac 1{2^n} <1\]
for each \(n=1,2,3,\dotsc\).
Some evidence:
But how can we prove this in general?
Consider a row of dominoes.
Suppose that two conditions are satisfied:
Then all dominoes fall eventually.
Consider a sequence
\[\mathrm{Claim}(1), \mathrm{Claim}(2), \dotsc\]
of claims. (That is, each \(\mathrm{Claim}(n)\) is either true or false.)
Suppose that two conditions are satisfied:
Then \(\mathrm{Claim}(n)\) is true for all natural numbers \(n=1,2,\dotsc\)
The principle of mathematical induction is one of the crucial properties that characterise what natural numbers are.
Further reading: Peano axioms
Let's use mathematical induction to prove that
\[1 + 2 + \dotsb + n = \frac{n(n+1)}2 \qquad \left( = \binom{n+1}2\right)\]
for each \(n = 1,2,\dotsc\)
We take \(\mathrm{Claim}(n)\) to be the statement \(1+2+\dotsb+n = \frac{n(n+1)}2\).
Base case:
\(1 = \frac{1 (1+1)}2\) so \(\mathrm{Claim}(1)\) is true.
Induction step:
Suppose that \(\mathrm{Claim}(n)\) is true for some \(n\).
That is, \(1 + 2 + \dotsb + n = \frac{n(n+1)}2\).
We seek to show that \(\mathrm{Claim}(n+1)\) is also true.
Let's compute:
\[\begin{aligned}\text{LHS of } \mathrm{Claim}(n+1) & = 1 + \dotsb + (n+1)\\&={(1+\dotsb + n)} + (n+1) \\ & = \frac{n(n+1)}2 + (n+1)\\ & = \frac{n(n+1)}2 + \frac{2n+2}2\\ & = \frac{n^2 + 3n+2}2 \\& = \frac{(n+1)(n+2)}2 = \text{RHS of }\mathrm{Claim}(n+1)\end{aligned}\]
Hence, \(\mathrm{Claim}(n+1)\) is true too.
We conclude that, by induction, \(\mathrm{Claim}(n)\) is true for all \(n=1,2,\dotsc\).
Problem: Produce a "closed formula" for
\[1^2 + \dotsb+ n^2.\]
Let's first compute some values:
\[\begin{array}{r|l}n & 1^2+\dotsb + n^2 \\ \hline 1 & 1 \\ \hline 2 & 5 = 1 + 4 \\ \hline 3 & 14 = 1 + 4 + 9 \\ \hline 4 & 30 = 14 + 16 \\ \hline 5 & 55 = 30 + 25 \\ \hline 6 & 91 = 55 + 36\end{array}\]
Inspiration, coffee, or Lagrange polynomials suggest that perhaps
\[1^2 + \dotsb + n^2 = \frac{n(n+1)(2n+1)}6.\]
Exercise: Use induction to prove this! ("Guess and verify")
Base case:
\(1^2 = \frac {1\cdotp 2 \cdotp 3}6\)
Induction step \(n \to n+1\):
\[\begin{aligned}1^2 + \dotsb + (n+1)^2 & = (1^2 + \dotsb + n^2) + (n^2 + 2n+1) \\ & = \frac{n(n+1)(2n+1)}6 + \frac{6n^2+12n+6}6\\&= \frac{2n^3 + 9n^2 + 13n+6}6 \\ & = \frac{(n+1)(n+2)(2(n+1)+1)}6\end{aligned}.\]
We could go on! For example,
\[1^3 + \dotsb + n^3 = \frac{n^2(n+1)^2}4 = (1^2 + \dotsb + n^2)^2.\]
General case: Faulhaber's formula for \(1^p + \dotsb + n^p\).
"Real-life" proofs by induction are often less straightforward than what we did so far.
For example, we may need to prove stronger claims (than what we're really after) for the induction step to work.
We may now use induction to finally prove that
\[ \frac 1 2 + \frac 1 4 + \dotsb + \frac 1 {2^n} < 1\]
for \(n = 1,2,\dotsc\).
\(\mathrm{Claim}(n)\colon \frac 1 2 + \dotsb + \frac 1{2^n} < 1\).
Base case: \(\mathrm{Claim}(1)\) is true because \(\frac 1 2 < 1\).
Induction step: Suppose that \(\mathrm{Claim}(n)\) is true. In other words, \(\frac 1 2 + \dotsb + \frac 1{2^n} < 1\). We need to show that \(\frac 1 2 + \dotsb + \frac 1{2^{n+1}} < 1\).
We compute:
\[\begin{aligned}\frac 1 2 + \dotsb + \frac 1{2^{n+1}} &= \underbrace{\frac 1 2 + \dotsb + \frac 1{2^n}}_{<1 \text{ by assumption}} + \frac 1{2^{n+1}}\\& < 1 + \frac 1{2^{n+1}}\end{aligned}.\]
Not good enough! What can we do?
Inspired by the sequence \(\frac 1 2, \frac 3 4, \frac 7 8, \dotsc\) we try to prove
\[\mathrm{Claim}'(n)\colon \frac 1 2 + \dotsb + \frac 1{2^n} = 1 - \frac 1 {2^n}.\]
Base case: \(\frac 1 2 = 1 - \frac 1 2\) so \(\mathrm{Claim}'(1)\) is true.
Induction step: Suppose that \(\frac 1 2 + \dotsb + \frac 1 {2^n} = 1 - \frac 1 {2^n}\). Then:
\[\begin{aligned}\frac 1 2 + \dotsb + \frac 1{2^{n+1}} &= {\frac 1 2 + \dotsb + \frac 1{2^n}} + \frac 1{2^{n+1}}\\& = 1 - \frac 1 {2^n} + \frac 1{2^{n+1}} = 1 - \frac 2{2^{n+1}} + \frac 1 {2^{n+1}}\\& = 1 - \frac 1 {2^{n+1}}.\end{aligned}.\]
Hence, \(\frac 1 2 + \dotsb + \frac 1 {2^n} = 1 - \frac 1{2^{n}} < 1\) for each \(n = 1,2,\dotsc\)
Show that
\[ 1 + \frac 1 {2^2} + \frac 1 {3^2} + \dotsb + \frac 1 {n^2} < 2\]
for each \(n = 1,2,\dotsc\)
We use induction to show that
\[ 1 + \frac 1 {2^2} + \frac 1 {3^2} + \dotsb + \frac 1 {n^2} \leqslant 2 - \frac 1 n\]
for each \(n = 1,2,\dotsc\)
Note that
\[\begin{aligned}&& 2-\frac 1 n + \frac 1{(n+1)^2} & \leqslant 2 - \frac 1 {n+1} \\&\Leftrightarrow&\frac 1 {n+1} + \frac 1{(n+1)^2} & \leqslant \frac 1 n\\&\Leftrightarrow&n(n+1)+n&\leqslant(n+1)^2\\&\Leftrightarrow&0 &\leqslant 1.\end{aligned}\]
A \(2\times 2\) board with one square removed is called an L-shape.
Consider tilings of boards using L-shapes (or other shapes such as "Tetris pieces") and subject to various restrictions.
Can we always tile an \(n\times n\) board (where \(n\geqslant 2\)) using L-shapes in such a way that there is precisely one square left?
Impossible! \(8 = 9-1\) is not divisible by 3.
Show that any \(2^n\times 2^n\) board can be tiled using L-shapes in a way that excludes just one square.
Draw finitely many straight lines (infinite in both directions) in the plane.
Show that the resulting regions can be coloured red and blue such that any two regions that touch more than one common point are coloured differently.
Draw finitely many straight lines (infinite in both directions) in the plane.
Show that the resulting regions can be coloured red and blue such that any two regions that touch more than one common point are coloured differently.
What is the number of subsets \(\{1,\dotsc,n\}\)?
Let \(x\not= 1\) and \(n\geqslant 0\).
Show (by induction or other means) that \[1 + x + x^2 + \dotsb + x^n = \frac{1-x^{n+1}}{1-x}.\]
Note that taking \(x = \frac 1 2\), we recover one of the first problems from today's session.
Find the number of compositions of \(n\).
A composition of an integer \(n\geqslant 1\) is a sequence of integers
\[(m_1,\dotsc,m_r)\]
with \(r,m_1,\dotsc,m_r\geqslant 1\) such that
\[n = m_1 + \dotsb + m_r.\]
\((3,1,2)\), \((2,4)\), and \((6)\) are examples of compositions of \(6\).
Hint. You'll want to use the following variant of induction, called strong induction:
Suppose that for all \(n\geqslant 1\), whenever \(\mathrm{Claim}(1),\dotsc,\mathrm{Claim}(n-1)\) are all true, then \(\mathrm{Claim}(n)\) is also true. Then \(\mathrm{Claim}(n)\) is true for all \(n\geqslant 1\).
The sequence \(f_0,f_1,\dotsc\) of Fibonacci numbers is defined via \(f_0 = 0\), \(f_1 = 1\), and \(f_n = f_{n-1} + f_{n-2}\) for \(n\geqslant 2\).
Show that:
Let \(n \geqslant 1\) be an integer and let \(x \geqslant -1\).
Then \[(1+x)^n \geqslant 1 + nx.\]
Define \(a_n = (1 + \frac 1 n)^n\) for \(n = 1,2,\dotsc\).
Show that \(a_1\leqslant a_2\leqslant \dotsb\).
Our sequence converges to Euler's number \(e=2.71828\dotso\).
Show that \(n! >2^n\) for all integers \(n > 3\).
Let \(a\) and \(b\) be any two integers and let \(n\) be a positive integer.
Show that \(a^n-b^n\) is divisible by \(a-b\).
Call a positive integer square-full if each of its prime factors occurs at least to a second power. Prove that there are infinitely many pairs of consecutive square-full numbers.
D. S. Gunderson. Handbook of mathematical induction. Theory and Applications. Chapters 1–3. CRC Press (2011)