Skip to main content
Contents Index
Search Book
Search Results:
No results.
Readability settings Prev Up Next
\(\require{cancel}
\newcommand{\nth}[1][n]{{#1}^{\mathrm{th}}}
\newcommand{\bbrac}[1]{\bigl(#1\bigr)}
\newcommand{\Bbrac}[1]{\Bigl(#1\Bigr)}
\newcommand{\correct}{\boldsymbol{\checkmark}}
\newcommand{\incorrect}{\boldsymbol{\times}}
\newcommand{\inv}[2][1]{{#2}^{-{#1}}}
\newcommand{\leftsub}[3][1]{\mathord{{}_{#2\mkern-#1mu}#3}}
\newcommand{\N}{\mathbb{N}}
\newcommand{\Z}{\mathbb{Z}}
\newcommand{\Q}{\mathbb{Q}}
\newcommand{\R}{\mathbb{R}}
\newcommand{\I}{\mathbb{I}}
\newcommand{\abs}[1]{\left\lvert #1 \right\rvert}
\DeclareMathOperator{\sqrtop}{sqrt}
\newcommand{\lgcnot}{\neg}
\newcommand{\lgcand}{\wedge}
\newcommand{\lgcor}{\vee}
\newcommand{\lgccond}{\rightarrow}
\newcommand{\lgcbicond}{\leftrightarrow}
\newcommand{\lgcimplies}{\Rightarrow}
\newcommand{\lgcequiv}{\Leftrightarrow}
\newcommand{\lgctrue}{\mathrm{T}}
\newcommand{\lgcfalse}{\mathrm{F}}
\newcommand{\boolnot}[1]{{#1}'}
\newcommand{\boolzero}{\mathbf{0}}
\newcommand{\boolone}{\mathbf{1}}
\newcommand{\setdef}[2]{\left\{\mathrel{}#1\mathrel{}\middle|\mathrel{}#2\mathrel{}\right\}}
\newcommand{\inlinesetdef}[2]{\{\mathrel{}#1\mathrel{}\mid\mathrel{}#2\mathrel{}\}}
\let\emptyword\emptyset
\renewcommand{\emptyset}{\varnothing}
\newcommand{\relcmplmnt}{\smallsetminus}
\newcommand{\union}{\cup}
\newcommand{\intersection}{\cap}
\newcommand{\cmplmnt}[1]{{#1}^{\mathrm{c}}}
\newcommand{\disjunion}{\sqcup}
\newcommand{\cartprod}{\times}
\newcommand{\words}[1]{{#1}^\ast}
\newcommand{\length}[1]{\abs{#1}}
\newcommand{\powsetbare}{\mathcal{P}}
\newcommand{\powset}[1]{\powsetbare(#1)}
\newcommand{\funcdef}[4][\to]{#2\colon #3 #1 #4}
\newcommand{\ifuncto}{\hookrightarrow}
\newcommand{\ifuncdef}[3]{\funcdef[\ifuncto]{#1}{#2}{#3}}
\newcommand{\sfuncto}{\twoheadrightarrow}
\newcommand{\sfuncdef}[3]{\funcdef[\sfuncto]{#1}{#2}{#3}}
\newcommand{\funcgraphbare}{\Delta}
\newcommand{\funcgraph}[1]{\funcgraphbare(#1)}
\newcommand{\nmathrel}[1]{\mathrel{\not #1}}
\newcommand{\relset}[3]{#1_{{} #2 #3}}
\newcommand{\gtset}[2]{\relset{#1}{\gt}{#2}}
\newcommand{\posset}[1]{\gtset{#1}{0}}
\newcommand{\geset}[2]{\relset{#1}{\ge}{#2}}
\newcommand{\nnegset}[1]{\geset{#1}{0}}
\newcommand{\neqset}[2]{\relset{#1}{\neq}{#2}}
\newcommand{\nzeroset}[1]{\neqset{#1}{0}}
\newcommand{\ltset}[2]{\relset{#1}{\lt}{#2}}
\newcommand{\leset}[2]{\relset{#1}{\le}{#2}}
\newcommand{\natnumlt}[1]{\ltset{\N}{#1}}
\DeclareMathOperator{\id}{id}
\newcommand{\inclfunc}[2]{\iota_{#1}^{#2}}
\newcommand{\projfunc}[1]{\rho_{#1}}
\DeclareMathOperator{\proj}{proj}
\newcommand{\funcres}[2]{\left.{#1}\right\rvert_{#2}}
\newcommand{\altfuncres}[2]{\left.{#1}\right\rvert{#2}}
\DeclareMathOperator{\res}{res}
\DeclareMathOperator{\flr}{flr}
\newcommand{\floor}[1]{\lfloor {#1} \rfloor}
\newcommand{\funccomp}{\circ}
\newcommand{\funcinvimg}[2]{\inv{#1}\left({#2}\right)}
\newcommand{\card}[1]{\left\lvert #1 \right\rvert}
\DeclareMathOperator{\cardop}{card}
\DeclareMathOperator{\ncardop}{\#}
\newcommand{\EngAlphabet}{\{ \mathrm{a}, \, \mathrm{b}, \, \mathrm{c}, \, \dotsc, \, \mathrm{y}, \, \mathrm{z} \}}
\newcommand{\ShortEngAlphabet}{\{ \mathrm{a}, \, \mathrm{b}, \, \dotsc, \, \mathrm{z} \}}
\newcommand{\eqclass}[1]{\left[#1\right]}
\newcommand{\partorder}{\preceq}
\newcommand{\partorderstrict}{\prec}
\newcommand{\npartorder}{\npreceq}
\newcommand{\subgraph}{\preceq}
\newcommand{\subgraphset}[1]{\mathcal{S}(#1)}
\newcommand{\connectedsubgraphset}[1]{\mathcal{C}(#1)}
\newcommand{\permcomb}[3]{{#1}(#2,#3)}
\newcommand{\permcombalt}[3]{{#1}^{#2}_{#3}}
\newcommand{\permcombaltalt}[3]{{\leftsub{#2}{#1}}_{#3}}
\newcommand{\permutation}[2]{\permcomb{P}{#1}{#2}}
\newcommand{\permutationalt}[2]{\permcombalt{P}{#1}{#2}}
\newcommand{\permutationaltalt}[2]{{\permcombaltalt{P}{#1}{#2}}}
\newcommand{\combination}[2]{\permcomb{C}{#1}{#2}}
\newcommand{\combinationalt}[2]{\permcombalt{C}{#1}{#2}}
\newcommand{\combinationaltalt}[2]{{\permcombaltalt{C}{#1}{#2}}}
\newcommand{\choosefuncformula}[3]{\frac{#1 !}{#2 ! \, #3 !}}
\DeclareMathOperator{\matrixring}{M}
\newcommand{\uvec}[1]{\mathbf{#1}}
\newcommand{\zerovec}{\uvec{0}}
\newcommand{\lt}{<}
\newcommand{\gt}{>}
\newcommand{\amp}{&}
\definecolor{fillinmathshade}{gray}{0.9}
\newcommand{\fillinmath}[1]{\mathchoice{\colorbox{fillinmathshade}{$\displaystyle \phantom{\,#1\,}$}}{\colorbox{fillinmathshade}{$\textstyle \phantom{\,#1\,}$}}{\colorbox{fillinmathshade}{$\scriptstyle \phantom{\,#1\,}$}}{\colorbox{fillinmathshade}{$\scriptscriptstyle\phantom{\,#1\,}$}}}
\)
Section 7.1 Principle of Mathematical Induction
Axiom 7.1.1 . Principle of Mathematical Induction.
Suppose \(P(n) \) is a predicate where the variable \(n \) has domain the positive, whole numbers. If
\(P(1) \) is true, and
\((\forall k)\bbrac{P(k) \lgccond P(k+1)} \) is true,
then \((\forall n)P(n) \) is true.
It is usual to take the principle of mathematical induction as an
axiom ; that is, we
assume that mathematical induction is valid without proving it.
Aside: A look ahead.
We will discuss axioms a little more in
ChapterΒ 8 .
Below is an outline of the idea behind why it is reasonable to assume that mathematical induction is valid. However, this outline does not constitute a proof since it technically uses mathematical induction implicitly.
Reasonableness of mathematical induction.
Suppose
\(n \) is fixed. We have a sequence of valid arguments:
\(P(1) \lgccond P(2) \)
\(P(1) \)
\(P(2) \)
\(P(2) \lgccond P(3) \)
\(P(2) \)
\(P(3) \)
\begin{equation*}
\cdots
\end{equation*}
\(P(n-1) \lgccond P(n) \)
\(P(n-1) \)
\(P(n) \)
Each is valid (modus ponens). So if we make the two assumptions stated in the principles (\(P(1) \) is true and \(P(k) \lgccond P(k+1) \) is always true) we can trace the flow of truth from premises to conclusion in each argument in turn:
First argument
Premises true so conclusion is true.
Second argument
Premises true (using conclusion of the first argument) so conclusion is true.
Third argument
\(\nth[(n-1)] \) argument
Premises true (using conclusion of the
\(\nth[(n-2)] \) argument) so conclusion is true.
The conclusion of \(\nth[(n-1)] \) argument is \(P(n) \text{,}\) so \(P(n) \) is true.
Now, here is some specific terminology associated to proofs by induction.
base case
the statement
\(P(1) \) in a proof by mathematical induction
induction step
the portion of a proof by mathematical induction that establishes the statement
\((\forall k) \bbrac{ P(k) \lgccond P(k+1) } \)
induction hypothesis
the assumption
\(P(k) \) made as the first step in the induction step of a proof by mathematical induction
Procedure 7.1.2 . Proof by induction.
To prove a universal statement indexed by whole numbers:
Base case.
Start by proving the statement obtained from the universally quantified predicate for the base case
\(n = 1 \text{.}\)
Induction step.
Next, assume that
\(k \) is a fixed number such that
\(k \ge 1 \text{,}\) and that the statement obtained from the universally quantified predicate is true for
\(n = k \text{.}\) Based on this assumption, try to prove that the next case,
\(n=k+1 \text{,}\) is also true.
Worked Example 7.1.3 .
Prove that the sum of the first \(n \) positive integers is
\begin{equation*}
\frac{n(n+1)}{2}\text{.}
\end{equation*}
Solution .
We want to prove \((\forall n)P(n) \text{,}\) where \(P(n) \) is as follows.
\begin{gather*}
P(1) \colon \:\: 1 = \frac{1\cdot 2}{2}, \qquad
P(2) \colon \:\: 1 + 2 = \frac{2 \cdot 3}{2}, \qquad
P(3) \colon \:\: 1 + 2 + 3 = \frac{3 \cdot 4}{2},\\
\dotsc, \qquad
P(n) \colon \:\: 1 + 2 + \dotsb + n = \frac{n(n+1)}{2}, \qquad
\dotsc
\end{gather*}
We will prove this by induction.
Base case.
\(1 = (1\cdot2)/2 \) is obviously true.
Induction step.
Assume the statement is true for \(n=k \text{;}\) that is, assume that
\begin{equation*}
1 + 2 + \dotsb + k = k(k+1)/2 \text{.}
\end{equation*}
We want to show that this implies the statement is true for \(n = k+1 \text{;}\) that is, to show
\begin{equation*}
1 + 2 + \dotsb + k + (k + 1) = (k + 1) (k + 2) / 2 \text{.}
\end{equation*}
We have
\begin{align*}
1 + 2 + \dotsb + k + (k + 1) \amp = (1 + 2 + \dotsb + k) + (k + 1) \\
\amp = \frac{k (k + 1)}{2} + (k + 1) \\
\amp = \frac{k^2 + 3 k + 2}{2} \\
\amp = \frac{(k + 1)(k + 2)}{2} \text{.}
\end{align*}
Worked Example 7.1.4 .
Prove that
\(n^3 + (n+1)^3 + (n+2)^3 \) is always divisible by
\(9 \) for every
\(n \ge 1 \text{.}\)
Solution .
We want to prove \((\forall n)P(n) \text{,}\) where \(P(n) \) is as follows.
\begin{align*}
P(1) \amp \colon \quad 1^3 + 2^3 + 3^3 \text{ is divisible by } 9 \\
P(2) \amp \colon \quad 2^3 + 3^3 + 4^3 \text{ is divisible by } 9 \\
P(3) \amp \colon \quad 3^3 + 4^3 + 5^3 \text{ is divisible by } 9 \\
\amp \vdots \\
P(n) \amp \colon \quad n^3 + (n+1)^3 + (n+2)^3 \text{ is divisible by } 9 \\
\amp \vdots
\end{align*}
We will prove this by induction.
Base case.
For
\(n=1 \text{,}\) \(n^3 + (n+1)^3 + (n+2)^3 = 36 = 9\cdot 4 \text{.}\)
Induction step.
Assume \(k^3 + (k+1)^3 + (k+2)^3 \) is divisible by \(9 \text{.}\) This means that there exists some whole number \(m \) so that
\begin{equation*}
k^3 + (k+1)^3 + (k+2)^3 = 9 m \text{.}
\end{equation*}
We want to show that \((k+1)^3 + (k+2)^3 + (k+3)^3 \) is also divisible by \(9 \text{.}\) To make the connection between this sum of cubes and the βprevious caseβ sum of cubes above, we can add in (and simultaneously subtract out) a \(k^3 \) term:
\begin{align*}
(k+1)^3 + (k+2)^3 + (k+3)^3 \amp = \bbrac{k^3 + (k+1)^3 + (k+2)^3} + (k+3)^3 - k^3 \\
\amp = 9m + (k+3)^3 - k^3 \\
\amp = 9m + (k^3 + 9k^2 + 27k + 27) - k^3 \\
\amp = 9(m + k^2 + 3k + 3) \text{.}
\end{align*}
Since we have factored our sum of cubes into a product involving \(9 \text{,}\) that sum of cubes is divisible by \(9 \text{.}\)
Worked Example 7.1.5 .
Prove that
\(3^{3n} + 1 \) is divisible by
\(7 \) whenever
\(n \) is odd.
Solution .
We do not want to use \(n \) as our induction index, since it jumps by twos. But \(n \) odd means that \(n = 2m - 1 \) for some \(m\ge 1 \text{,}\) so we want to prove \((\forall m)P(m) \text{,}\) where \(P(m) \) is as follows.
\begin{align*}
P(1) \amp \colon \quad 3^3 + 1 \text{ is divisible by } 7 \\
P(2) \amp \colon \quad 3^9 + 1 \text{ is divisible by } 7 \\
P(3) \amp \colon \quad 3^{15} + 1 \text{ is divisible by } 7 \\
\amp \vdots \\
P(m) \amp \colon \quad 3^{6m-3} + 1 \text{ is divisible by } 7 \\
\amp \vdots
\end{align*}
We will prove this by induction.
Base case.
For
\(m=1 \text{,}\) \(3^3+1 = 28 = 7\cdot 4 \text{.}\)
Induction step.
Assume \(3^{6k-3} + 1 \) is divisible by \(7 \text{.}\) This means that there exists some whole number \(\ell \) so that
\begin{equation*}
3^{6k-3} + 1 = 7 \ell \text{.}
\end{equation*}
We want to show \(3^{6(k+1)-3} + 1 \) is divisible by \(7 \text{.}\) We have
\begin{align*}
3^{6 (k + 1) - 3} + 1 \amp = (3^{6k - 3}) (3^6) + 1 \\
\amp = (3^{6k-3} + 1 - 1) (3^6) + 1 \\
\amp = (3^{6k-3} + 1) (3^6) - 3^6 + 1 \\
\amp = (7 \ell) (3^6) - 728 \\
\amp = 7 (3^6 \ell - 104) \text{.}
\end{align*}
Since we have our expression factored into a product involving \(7 \text{,}\) our expression is divisible by \(7 \) as desired.
Worked Example 7.1.7 .
Prove
\(2^n \lt n! \) whenever
\(n \ge 4 \text{.}\)
Solution .
Base case.
For
\(n=4 \text{,}\) \(2^4 = 16 \) and
\(4! = 24 \text{.}\)
Induction step.
Assume \(2^k \lt k! \) for some \(k \ge 4 \text{.}\) We want to show \(2^{k+1} \lt (k+1)! \text{.}\) We have
\begin{equation*}
2^{k+1} = 2(2^k) \lt 2(k!) \lt (k+1)(k!) = (k+1)! \text{.}
\end{equation*}
Aside: Note.
As you can easily check, the statement in
Worked ExampleΒ 7.1.7 is actually false for
\(n = 1,2,3 \text{.}\)