MathLinks Forum LaTeX Help AoPS Classes Books Classroom MathLinks Contest Math Resources
The time now is Thu Mar 11, 2010 5:21 pm
All times are UTC + 2
View posts since last visit
View unanswered posts
View previous topicView next topic

Moderators: tetrahedr0n
 
4 Posts • Page 1 of 1
Author Message
ZennyK
Riemann Hypothesis
Riemann Hypothesis


Offline
Joined: 16 Jun 2004
Posts: 355
Location: Working on the Riemann Hypothesis
United States

To rate posts you must be logged in
#1
Year 15, Round 4, Problem 3

Since I have no idea when the solutions will be online, could someone please post their solution to problem 3 of round 4, or at least give the answer? Thank you in advance whoever replies to this.
_________________
"If chance be the Father of all flesh, then disaster is his rainbow in the sky." - Steve Turner

PostPosted: Tue Jun 22, 2004 9:26 pm
JBL
Birch & Swinnerton Dyer
Birch & Swinnerton Dyer

Offline
Joined: 05 Jul 2003
Posts: 11277
Location: Brooklyn, NY or Cambridge, MA
United States

To rate posts you must be logged in
#2
The answer is any variant on the form:

Click to reveal hidden content
(x - y)(x - z)(x - w)(y - z)(y - w)(z - w).

_________________
Joel
Hi Deeps! <3

PostPosted: Tue Jun 22, 2004 10:06 pm
Yasha
Riemann Hypothesis
Riemann Hypothesis


Offline
Joined: 27 May 2004
Posts: 255
Location: Pasadena, CA

To rate posts you must be logged in
#3
I'd post my solution, but I'm at MOP now and my solution is at home. I do remember that I did get the same polynomial as JBL.
_________________
--Yasha B.-K.

PostPosted: Wed Jun 23, 2004 1:30 am
Erin J. Schram
Poincare Conjecture
Poincare Conjecture

Offline
Joined: 01 Jun 2004
Posts: 204
Location: Savage, Maryland
United States

To rate posts you must be logged in
#4
Re: Year 15, Round 4, Problem 3

ZennyK wrote:
Since I have no idea when the solutions will be online, could someone please post their solution to problem 3 of round 4, or at least give the answer? Thank you in advance whoever replies to this.


I had written up a credits and quick answers page for round 4 months in advance. I don't know why it never got posted at the NSA web site.

I have the LaTeX file for it. So I will post it here. I will also enclose it in spoiler brackets even though Year 15 Round 4 is over. I hope you can read raw LaTeX.

Erin Schram

Click to reveal hidden content
\documentclass[12pt]{article}
%\pagestyle{empty}
\setlength{\oddsidemargin}{0pt}
\setlength{\evensidemargin}{0pt}
\setlength{\topmargin}{-0.5in}
\setlength{\textwidth}{6.5in}
\setlength{\textheight}{9in}
\usepackage[pdftex]{graphicx}
\usepackage{amsmath,amssymb}
\begin{document}
\begin{center}\large\bfseries
USA Mathematical Talent Search\\
CREDITS and QUICK ANSWERS\\
Round 4 --- Year 15 --- Academic Year 2003--2004\\
\end{center}
\setlength{\parindent}{0pt}
\setlength{\parskip}{3pt}

\textbf{1/4/15.}
\parbox[t]{5.75in}{\footnotesize\sffamily
Find, with proof, the smallest positive integer $n$ for which the sum of the digits of $29n$ is as small as possible.}

This problem was fashioned after Problem 3 of the 2nd Lithuanian Olympiad administered at Vilnius University on September 30, 2000.

We search for $29n$ rather than directly for $n$. Numbers with only a single nonzero digit are that digit times a power of ten, which would not be divisible by $29$. Of the numbers with two nonzero digits, the smallest digit sum occurs with both nonzero digits being~$1$. The smallest such number must begin and end in~$1$, because if it ended in zero, we could divide by $10$. So we are looking for the smallest integer of the form $10^e+i$ that is a multiple of $29$.

The powers of $10$ mod~$29$ are $1$, $10$, $13$, $14$, $24$, $8$, $22$, $17$, $25$, $18$, $6$, $2$, $20$, $26$, $28$, $19$, $16$, $15$, $5$, $21$, $7$, $12$, $4$, $11$, $23$, $27$, $9$, $3$, $1$, $\ldots$. so $10^{14}\equiv 28\pmod{29}$. Therefore $10^{14}+1$ is divisible by $29$, and $n=(10^{14}+1)/29 = 3448275862069$.
\bigskip

\textbf{2/4/15.}
\parbox[t]{5.75in}{\footnotesize\sffamily
For four integer values of $n$ greater than six, there exist right triangles whose side lengths are integers equivalent to $4$, $5$, and $6$ modulo~$n$, in some order. Find those values. Prove that at most four such values exist. Also, for at least one of those values of $n$, provide an example of such a triangle.}

This problem was devised by Dr.~Peter Anspach of the NSA, after a lunchtime conversation among mathematicians about trying to create a $(4,5,6)$ right triangle.

Let $a$, $b$, and $c$ be the sides of such a right triangle, with $a^2+b^2=c^2$. If $c\equiv 4\pmod n$, then $5^2+6^2\equiv 4^2\pmod n$ so $n$ is a factor of $5^2+6^2-4^2 = 45$. If $c\equiv 5\pmod n$, then $n$ is a factor of $4^2+6^2-5^2 = 27$. If $c\equiv 6\pmod n$, then $n$ is a factor of $4^2+5^2-6^2 = 5$, but since $n>6$, we can reject this case. Thus, $n$ is a factor of $27$ or $45$, so $n$ is $9$, $15$, $27$, or $45$.

Given an $n$, a triple $(a,b,c)$ that forms a right triangle can be found by trial and error, by applying number theory to an equation such as $(9x+5)^2+(9y+6)^2 =(9z+4)^2$, or by the $(r(s^2-t^2),2rst,r(s^2+t^2))$ method of finding Pythagorean triples after solving for $r$, $s$, and $t$ that give the right values mod~$n$. The first three solutions for each case are:
\begin{description}
\item for $n=9$ and $c\equiv 4\pmod{9}$: $(24,32,40)$, $(51,68,85)$, $(78,104,130)$;
\item for $n=9$ and $c\equiv 5\pmod{9}$: $(40,96,104)$, $(15,112,113)$, $(60,175,185)$;
\item for $n=15$ and $c\equiv 4\pmod{15}$: $(65,156,169)$, $(140,336,364)$, $(215,516,559)$;
\item for $n=27$ and $c\equiv 5\pmod{27}$: $(168,490,518)$, $(33,544,545)$, $(814,1248,1490)$;
\item for $n=45$ and $c\equiv 4\pmod{45}$: $(1176, 2975,3199)$, $(3920,16926,17374)$, $(1760,17556,17644)$.
\end{description} \vfil
\eject

\textbf{3/4/15.} \parbox[t]{5.75in}{\footnotesize\sffamily Find a nonzero polynomial $f(w,x,y,z)$ in the four indeterminates $w$, $x$, $y$, and~$z$ of minimum degree such that switching any two indeterminates in the polynomial gives the same polynomial except that its sign is reversed. For example, $f(z,x,y,w) = -f(w,x,y,z)$. Prove that the degree of the polynomial is as small as possible.}

This problem was invented by Dr.~David Grabiner, an NSA mathematician who helps with the evaluation and grading of problems for the USAMTS.

Consider any two of the indeterminants, say $w$ and $x$.
If the standard form of $f (w,x,y,z)$ has a term $kw^ax^by^cz^d$, then it must have a matching term $-kw^bx^ay^cz^d$. Grouping them together gives either $k(w^{a-b}-x^{a-b})w^bx^by^cz^d$ or $k(x^{b-a}-w^{b-a})w^ax^ay^cz^d$, which both are divisible by $w-x$. Since the
entire polynomial can be grouped this way, it is divisible by $w-x$.

Thus, for every pair of indeterminants, the polynomial is divisible by their difference. This means $f(w,x,y,z)$ is divisible by $(w-x)(w-y)(w-z)(x-y)(x-z)(y-z)$. Letting $f(w,x,y,z)= (w-x)(w-y)(w-z)(x-y)(x-z)(y-z)$ keeps the degree to a minimum, and it works because switching any two indeterminants in $(w-x)(w-y)(w-z)(x-y)(x-z)(y-z)$ gives the same polynomial except that
its sign is reversed.
\bigskip

\textbf{4/4/15.}
\parbox[t]{5.75in}{\footnotesize\sffamily
For each nonnegative integer $n$ define the function $f_n(x)$ by
\begin{displaymath}
f_n(x) = \sin^n(x) + \sin^n(x + \frac{2\pi}{3}) + \sin^n(x + \frac{4\pi}{3})
\end{displaymath}
for all real numbers $x$, where the sine functions use radians. The functions $f_n(x)$ can be also expressed as polynomials in $\sin(3x)$ with rational coefficients. For example, \begin{eqnarray*}
&\displaystyle f_0(x) = 3,\qquad f_1(x) = 0,\qquad f_2(x) = \frac32,\qquad f_3(x) = -\frac34\sin(3x),\\
\noalign{\smallskip}
&\displaystyle f_4(x) = \frac98,\qquad f_5(x) = -\frac{15}{16}\sin(3x),\qquad f_6(x) = \frac{27}{32} + \frac{3}{16}\sin^2(3x),
\end{eqnarray*}
for all real numbers~$x$. Find an expression for $f_7(x)$ as a polynomial in $\sin(3x)$ with rational coefficients, and prove that it holds for all real numbers $x$.}

This problem was devised by the distinguished Hungarian mathematician and poet Mih\'aly Bencze of Brasso, Transylvania.

The triple angle formula for sine is $\sin(3x)=3\sin(x)-4\sin^3(x)$. But
since $\sin(3x)$ also equals $\sin(3({x + \frac{2\pi}{3}}))$ and $\sin(3({x + \frac{4\pi}{3}}))$, it also expands to $3\sin({x + \frac{2\pi}{3}}) - 4\sin^3({x + \frac{2\pi}{3}})$ and $3\sin({x + \frac{4\pi}{3}}) - 4\sin^3({x + \frac{4\pi}{3}})$. Thus, for all real numbers~$x$,
\begin{eqnarray*}
f_n(x)\sin(3x) &=& \sin^n(x)\sin(3x) + \sin^n(x + \frac{2\pi}{3})\sin(3x) + \sin^n(x + \frac{4\pi}{3})\sin(3x)\\
&=& \sin^n(x)\bigl(3\sin(x) - 4\sin^3(x)\bigr)\\
&& \qquad {}+ \sin^n(x + \frac{2\pi}{3})\bigl(3\sin(x + \frac{2\pi}{3}) - 4\sin^3(x + \frac{2\pi}{3})\bigr)\\
&& \qquad {}+ \sin^n(x + \frac{4\pi}{3})\bigl(3\sin(x + \frac{4\pi}{3}) - 4\sin^3(x + \frac{4\pi}{3})\bigr)\\
&=&3\bigl(\sin^{n+l}(x)+\sin^{n+l}(x + \frac{2\pi}{3})+\sin^{n+l}(x + \frac{4\pi}{3})\bigr)\\
&&\qquad {}-4\bigl(\sin^{n+3}(x) + \sin^{n+3}(x + \frac{2\pi}{3}) + \sin^{n+3}(x + \frac{4\pi}{3})\bigl)\\
&=&3f_{n+l}(x) - 4f_{n+3}(x).
\end{eqnarray*}
Setting $n$ to 4 and rearranging gives
$f_7(x) = \frac34f_5(x) - \frac14f_4(x)\sin(3x) = (\frac34)(-\frac{15}{16})\sin(3x) - (\frac14)(\frac98)\sin(3x) = (-\frac{63}{64})\sin(3x)$.
\bigskip

\textbf{5/4/15.} \parbox[t]{2.75in}{\footnotesize\sffamily
Triangle $ABC$ is an obtuse isosceles triangle with the property that three squares of equal size can be inscribed in it as shown on the right. The ratio $AC/AB$ is an irrational number that is the root of a cubic polynomial. Determine that polynomial.} \raisebox{-70pt}[0pt]{\setlength{\unitlength}{2pt}
\begin{picture}(95,40)
\put(10,6){\line(1,0){80}}
\put(10,6){\line(4,3){40}}
\put(90,6){\line(-4,3){40}}
\put (5,0){A}
\put (48,37){B}
\put(91,0){C}
\put(39.1,6){\line(0,1){21.8}}
\put(39.1,27.8){\line(1,0){21.8}}
\put(60.9,6){\line(0,1){21.8}}
\put(45.7,6){\line(-3,4){12.85}}
\put (45.7, 6){\line (4, 3){17.1}}
\put(50,36){\line(3,-4){12.85}}
\put(54.3,6){\line(3,4){12.85}}
\put(54.3,6){\line(-4,3){17.1}}
\put(50,36){\line(-3,-4){12.85}}
\put(29,24){D}
\put(36, 30){E}
\put(60,30){F}
\put(67,24){G}
\put(37,0){J}
\put(43,0){K}
\put (53,0){K}
\put(59,0){M}
\put (31,16){H}
\put(65.5,16){I}
\end{picture}}
\medskip

This triangle was apparently discovered by the Italian mathematician Eugenio Calabi. It is the only non-equilateral triangle into which we can fit three equal squares in this manner. The triangle is displayed on page 266 of {\em The Book of Numbers,} by John H. Conway and Richard K. Guy, published by Springer-Verlag in 1996.
\medskip

Let $s$ be the length of a side of the squares. By symmetry, $AJ=MC$, so $2(AJ)={AC-JM}={AC-s}$. $AD=AB-BD=AB-s$. Triangle $AEJ$ is congruent to triangle $AKD$, so $AD=AJ$. So ${AC-s}={2(AB-s)}$, giving $s=2(AB)-AC$ and $AJ=AC-AB$.

Let $P$ be the midpoint of side $\overline{AC}$. Triangle $AEJ$ is
similar to triangle $ABP$. So $AE/AJ=AB/AP=2(AB/AC)$. $AE=2(AC-AB)(AB/AC)$.

By the Pythagorean Theorem $(AE)^2=(AJ)^2+s^2$, which converts to \begin{eqnarray}
\left(2(AC-AB)\left(\frac{AB}{AC}\right)\right)^2
&=& (AC-AB)^2 + (2(AB)-AC)^2\\
\left(2(AC-AB)(AB)\right)^2
&=&(AC)^2(AC-AB)^2 + (AC)^2(2(AB)-AC)^2.
\end{eqnarray}
Let $x=AC/AB$. Equation (2) converts to
\begin{eqnarray*}
4(x-1)^2 &=& x^2(x-1)^2 + x^2(2-x)^2\\
(4-x^2)(x-1)^2 - x^2(2-x)^2 &=& 0\\
(2-x)(x+2)(x-1)^2 - x^2(2-x)^2 &=& 0\\
(2-x)\bigl((x+2)(x-1)^2 - x^2(2-x)\bigr) &=& 0\\
(2-x)(2x^3-2x^2-3x+2) &=& 0
\end{eqnarray*}
By the triangle inequality, $AC<AB+BC=2(AB)$, so $x<2$ and $2-x\ne 0$. Therefore, $AC/AB$ is the root of the cubic polynomial $2x^3-2x^2-3x+2$.

\end{document}


The End

PostPosted: Wed Jun 23, 2004 1:50 am
Display posts from previous:   Sort by:   
4 Posts • Page 1 of 1
View previous topicView next topic
Jump to:  

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum
You cannot post calendar events in this forum

Created and Maintained by Valentin Vornicu - (c) AoPS Inc. 2004-2008