\documentstyle[12pt,twoside,macpsfig]{article}
\input{macros-course}

\newenvironment{proposition*}{{\bf Proposition: }\em}{}
\newenvironment{Proof}{\proof }{\qed}


\begin{document}

\handout{ps1}{Problem Set 1}{February 1, 1994}
\pset{1}
{\bf Reading:\ }Maurer \& Ralston: Section 0.7, Chapter 2, Sections 7.1--3,
and Sections 7.5-7.6.



\medskip




The following oral problems are due in tutorial on Monday, February 7.
Some of the problems ask you to ``verify or falisfy'' a proof.  For
these problems, you must either justify the correctness of each step
of the proof, or else locate and explain the error.  It is not
sufficient to merely show that an incorrect proof is indeed wrong.
You must explain where the bug is.

\medskip

\begin{problems}

\oproblem Verify or falisify the following proof that if $a$ and $b$
are two equal real numbers, then $a=0$.
\begin{Proof}
\begin{eqnarray*}
a&=&b \\
a^2&=&ab \\
a^2-b^2&=&ab-b^2 \\
(a-b)(a+b)&=&(a-b)b \\
a+b&=&b \\
a&=&0 
\end{eqnarray*}
\end{Proof}

\oproblem
Verify or falsify the following syllogism.
\[
\begin{array}{ll}
&\hbox{Nothing is better than complete happiness.}\cr
&\hbox{A ham sandwich is better than nothing.}\cr
&\hbox to 4in {\hrulefill}\cr
%\therefore
 & 
\hbox{A ham sandwich is better than complete happiness.}\cr
\end{array}\]

\oproblem Consider the $2\times2$ integer lattice of all points in
$\integers\times\integers$. Given any nine points in this lattice, can
two of them be selected so that the midpoint of the line between them
lies in $\integers\times\integers$?

\oproblem Verify or falsify the following proof that there exist two
irrational numbers $a$ and $b$ such that $a^b$ is rational.
\begin{Proof}
Recall that $\sqrt{2}$ is irrational.  Suppose that
$\paren{\sqrt{2}}^{\sqrt{2}}$ is rational.  Then, let $a=\sqrt{2}$,
let $b=\sqrt{2}$, and the proof is finished.  Suppose, however, that
$\paren{\sqrt{2}}^{\sqrt{2}}$ is irrational.  Then, let
$a=\paren{\sqrt{2}}^{\sqrt{2}}$ and let $b=\sqrt{2}$.  Thus, we have
$a^b=\paren{\sqrt{2}^{\sqrt{2}}}^{\sqrt{2}}=
\paren{\sqrt{2}}^{\sqrt{2}\cdot\sqrt{2}} = \paren{\sqrt{2}}^2 = 2$,
which is rational, and we're done.
\end{Proof}

\oproblem
Prove that $\frac{n^7}{7}+\frac{n^3}{3}+\frac{11n}{21}$ is an integer.
(\hint Use the polynomial expansion of $(n+1)^7$.)

\oproblem Verify or falisfy the following proof that for $x\neq 0$, we
have $x^n=1$.
\begin{Proof}
We use induction on~$n$.  For the base step, observe that if $x \neq
0$, we have $x^0 = 1$.  We now assume that for all $k\leq n$, we have $x^k=0$.
We shall show that $x^{n+1}=0$.  By the induction hypothesis, we have
$x^n = x^{n-1} = 1$.  Hence, it follows that 
\begin{eqnarray*}
x^{n+1} & = & \frac{x^nx^n}{x^{n-1}}\\
& = &\frac{1\times 1}{1}\\
& = & 1\ . 
\end{eqnarray*}
\end{Proof}

\oproblem For any integer $n\ge 3$, consider a simple polygon of $n$
vertices drawn in the plane.  (A polygon is \defi{simple} if it does
not cross itself.)  An \defi{$n$-gon} is a polygon with $n$ vertices,
A {\em triangulation} of an $n$-gon $P$ is a partition of the interior
of $P$ into triangles such that the only vertices of $P$ that lie on
the boundary of any triangle in the triangulation are the triangles
vertices.

\begin{problemparts}

\problempart Prove that any triangulation of an $n$-gon is comprised
of exactly $n-2$ triangles.

\problempart Using the fact the sum of the interior angles of a
triangle is $\pi$, give a formula for the sum of the angles of an
$n$-gon.  Use induction to prove that your formula is correct.

\end{problemparts}

\oproblem Verify or falsify the following proof that $\sum_{i=1}^n i$
is less than or equal to some positive constant times~$n$.
\begin{Proof}
The proof is by induction on $n$. 

If $n=1$, then the summation evaluates to $1$, which can be bounded
above by any constant constant greater than $1$ times~$1$.

Assume that the statement is true for $n=k$, that is, that
\[\sum_{i=1}^k i \le ck\] for some constant~$c>1$.  Adding $k+1$ to both 
sides, we obtain
\begin{eqnarray*}
\sum_{i=1}^{k+1} i & \le & ck + (k+1)\\
& \le & (c+1)(k+1)\ .
\end{eqnarray*}
Since $c$ is a constant greater than 1, so is~$c+1$.  For $n=k+1$,
therefore, the sum $\sum_{i=1}^{n} i$ is also bounded above by a
constant times~$n$.  Consequently, by induction this property holds
for all natural numbers.
\end{Proof}

\oproblem\label{prob:angles}
Verify (or find a flaw in) the following proof that
there exists an obtuse angle with measure equal to that of a right angle.
You may find Figure~\ref{fig:angles} useful.

\begin{figure}[t]
\centerline{\psfig{figure=figures/fudged.ps,height=2in}}
\figcaption{A sketch of the geometric construction in
    \hbox{Problem~\ref{prob:angles}}.}
\label{fig:angles}
\end{figure}

\begin{proof}
Construct a rectangle $abcd$ labeled so that $\overline{ad}$ is a diagonal.
Construct $\overline{ae}$ such that $\angle eac$ has small measure. 
Draw $\overline{ed}$. Construct perpendicular bisectors to $\overline{ed}$
and to $\overline{cd}$. Let $h$ be the intersection of these bisectors.
Since the angle $\angle eac$ is small, $h$ will not lie between $\overline{cd}$
and
 $\overline{ab}$. Construct $\overline{ah}$, $\overline{bh}$, $\overline{eh}$,
 and $\overline{dh}$. Since
the line $\overline{fh}$ is a perpendicular bisect of $\overline{ed}$, 
 $\bigtriangleup ehd$ is 
isoceles, and hence the lengths of $\overline{eh}$ and $\overline{hd}$
 are equal. 
Similarly, since $abcd$ is a rectangle, $\overline{gh}$ is a perpendicular
bisector of $\overline{ab}$, so the  $\bigtriangleup ahb$ is isoceles. Therefore
 $\overline{ah}$ and $\overline{hb}$ are equal in length and 
$\angle bah$ is equal in measure to  $\angle hba$.

With these results, it becomes clear that $\bigtriangleup eah$ and
$\bigtriangleup dbh$ are congruent. So $\angle eah$ and $\angle dbh$
 have equal measure.
Subtracting the (equal) measures of  $\angle bah$ and $\angle abh$ respectively
yields that the measure of $\angle eab$ equals the measure of $\angle dba$.
\end{proof}

\oproblem
A full binary tree is a binary tree such that any node
has either 0 or 2 children. Prove that the number of 
interior nodes in a full binary tree is one less than
the number of leaves.


\bigskip


The following written problems are due 
at the beginning of class on Tuesday, February 8.
Remember to write up each problem on a separate page
or pages, with your name on each sheet.  
Read Handout~\ref{course-information} for the policy on problem sets.

\medskip

\wproblem
Recall that the binomial coefficients ${ n \choose k }$ 
(read $n$ choose $k$
are defined on the natural numbers by ${n \choose k}={n! \over (n-k)! k!}$.

\begin{problemparts}


\problempart
Justify the notation by proving
that the number of ways to divide $n$ distinct objects into
$k$ different sets is ${n \choose k}$.

\problempart
The binomial coefficients are precisely the numbers appearing in
Pascal's triangle. See Figure~\ref{fig:binom}.
Prove that ${n \choose k}= {n-1 \choose k-1} + {n-1 \choose k}$ for
$n\ge1$ and $0<k<n$. Use this formula to prove that (when appropriately
labeled) the $k^{\rm th}$ entry in row $n$ of Pascal's triangle is
${n \choose k}$. This standard labelling will be used for th
remainder of the problem.
\begin{figure}[t]
\[\begin{array}{ccccccccccccccccccccc}
&&&&&&&& 1 &&&&&&&&\cr
&&&&&&& 1&&1 &&&&&&&\cr
&&&&&& 1&&2&&1 &&&&&&\cr
&&&&& 1&&3&&3&&1 &&&&&\cr
&&&& 1&&4&&6&&4&&1 &&&&\cr
&&& 1&&5&&10&&10&&5&&1 &&&\cr
&& 1&&6&&15&&20&&15&&6&&1 &&\cr
& 1&&7&&21&&35&&35&&21&&7&&1 &\cr
\end{array}\]
\figcaption{The first eight rows of Pascal's triangle.
The first and last entries in the $n^{\rm th}$ row are required to be 1.
Any other entry is the sum of the two entries above it in the ${n-1}^{\rm st}$
row.}
\label{fig:binom}
\end{figure}

\problempart
Each row of Pascal's triangle increases until the middle and
then decreases. Prove this.

\problempart
Prove that $(x+y)^n=\sum_{k=0}^n {n \choose k} x^k y^{n-k}$.
Use this fact to show that the sum of the entries in the $n^{\rm th}$ row
of Pascal's triangle is $2^n$. Similarly, 
note that the alternating sum of entries
in any but the top row of Pascal's triangle is 0.

\problempart
Although the diagonals of Pascal's triangle are infinite, one
can take partial sum along them. The partial sum
of the first $n$ entries of the  $m^{\rm th}$ diagonal is 
${n+m \choose m+1}$. State this formally and prove it.

\problempart
It is possible to pick ``off-diagonal'' lines of numbers in Pascal's
triangle such that the sum of the numbers on the $n^{\rm th}$ line
is the $n^{\rm th}$ Fibonacci number. State this formally and prove it.

\end{problemparts}

\wproblem
\begin{problemparts}

\problempart
Suppose you are given an unlimited supply of 
two types of postage stamps, one with a 3 cent
denomination, the other with of 5 cent denomination.
Any sufficiently large amount of postage can be
made up using just these stamps. Exactly how large
does the postage have to be for this to work?
Prove it.

\problempart
Suppose instead you are give 4 cent and 6 cent
denominations. Any sufficiently large and even
amount of postage can be formed from these two
types of stamp. Again, determine how large the
postage has to be for this to work.

\problempart
Suppose you are given two types of stamps with
denominations $a$ and $b$ such that $a$ and $b$
are relatively prime. I.E. the greatest
common denominator of $a$ and $b$ is 1, that
is they have no factors (other than 1) in common.

Prove that any sufficiently large large amount
of postage, $n$, can be formed using only stamps
of denominations $a$ and $b$. In other words,
show that if $n$ is sufficiently large,
one can write $n=ax+by$ for two natural
numbers $x$ and $y$.

Conjecture and prove a formula $f(a,b)$
such that if $n\ge f(a,b)$ then $n$ can
be written as $ax+by$ as above. Is this
the best formula possible?

\problempart
Generalize your previous results to stamp
denominations $a$ and $b$ which are not
relatively prime.

\end{problemparts}


\end{problems}


\end{document}


