<?xml version="1.0" encoding="UTF-8"?>

<record version="16" id="29">
 <title>Collatz problem</title>
 <name>CollatzProblem</name>
 <created>2001-08-19 10:20:17</created>
 <modified>2003-11-05 16:17:52</modified>
 <type>Conjecture</type>
 <creator id="2" name="akrowne"/>
 <author id="2" name="akrowne"/>
 <classification>
	<category scheme="msc" code="11B37"/>
 </classification>
 <synonyms>
	<synonym concept="Collatz problem" alias="Ulam's Problem"/>
	<synonym concept="Collatz problem" alias="1-4-2 Problem"/>
	<synonym concept="Collatz problem" alias="Syracuse problem"/>
	<synonym concept="Collatz problem" alias="Thwaites conjecture"/>
	<synonym concept="Collatz problem" alias="Kakutani's problem"/>
	<synonym concept="Collatz problem" alias="3n+1 problem"/>
 </synonyms>
 <keywords>
	<term>Collatz</term>
	<term>Ulam</term>
 </keywords>
 <preamble>\usepackage{amssymb}
\usepackage{amsmath}
\usepackage{amsfonts}
\usepackage{graphicx}
\usepackage{xypic}</preamble>
 <content>We define the function $f : \mathbb{N} \longrightarrow \mathbb{N} $ (where $\mathbb{N}$ excludes zero) such that

$$ f(a) =  \left\{
\begin{array}{rl}
3a+1  &amp; \text{ if } a \text{ is odd }   \\
 a/2  &amp; \text{ if } a \text{ is even.}  
\end{array}
\right. $$

Then let the sequence $c_n$ be defined as $c_i = f(c_{i-1})$, with $c_0$ an arbitrary natural seed value.

It is conjectured that the sequence $c_0, c_1, c_2, \ldots$ will always end in  $1,4,2$, repeating infinitely.  This has been verified by computer up to very large values of $c_0$, but is unproven in general. It is also not known whether this problem is decideable.  This is generally called the \emph{Collatz problem}.

The sequence $c_n$ is sometimes called the ``hailstone sequence''.  This is because it behaves analogously to a hailstone in a cloud which falls by gravity and is tossed up again repeatedly.  The sequence similarly ends in an eternal oscillation.</content>
</record>
