852 lines
24 KiB
TeX
852 lines
24 KiB
TeX
\chapter{Combinatorics}
|
|
|
|
\index{combinatorics}
|
|
|
|
\key{Combinatorics} studies methods for counting
|
|
combinations of objects.
|
|
Usually, the goal is to find a way to
|
|
count the combinations efficiently
|
|
without generating each combination separately.
|
|
|
|
As an example, let's consider a problem where
|
|
our task is to calculate the number of representations
|
|
for an integer $n$ as a sum of positive integers.
|
|
For example, there are 8 representations
|
|
for the number $4$:
|
|
\begin{multicols}{2}
|
|
\begin{itemize}
|
|
\item $1+1+1+1$
|
|
\item $1+1+2$
|
|
\item $1+2+1$
|
|
\item $2+1+1$
|
|
\item $2+2$
|
|
\item $3+1$
|
|
\item $1+3$
|
|
\item $4$
|
|
\end{itemize}
|
|
\end{multicols}
|
|
|
|
A combinatorial problem can often be solved
|
|
using a recursive function.
|
|
In this case, we can define a function $f(n)$
|
|
that counts the number of representations for $n$.
|
|
For example, $f(4)=8$ according to the above example.
|
|
The function can be recursively calculated as follows:
|
|
\begin{equation*}
|
|
f(n) = \begin{cases}
|
|
1 & n = 1\\
|
|
f(1)+f(2)+\ldots+f(n-1)+1 & n > 1\\
|
|
\end{cases}
|
|
\end{equation*}
|
|
The base case is $f(1)=1$,
|
|
because there is only one way to represent the number 1.
|
|
Otherwise, we go through all possibilities for
|
|
the last number in the sum.
|
|
For example, in when $n=4$, the sum can end
|
|
with $+1$, $+2$ or $+3$.
|
|
In addition, we also count the representation
|
|
that only contains $n$.
|
|
|
|
The first values for the function are:
|
|
\[
|
|
\begin{array}{lcl}
|
|
f(1) & = & 1 \\
|
|
f(2) & = & 2 \\
|
|
f(3) & = & 4 \\
|
|
f(4) & = & 8 \\
|
|
f(5) & = & 16 \\
|
|
\end{array}
|
|
\]
|
|
It turns out that the function also has a closed-form formula
|
|
\[
|
|
f(n)=2^{n-1},
|
|
\]
|
|
which is based on the fact that there are $n-1$
|
|
possible positions for +-signs in the sum,
|
|
and we can choose any subset of them.
|
|
|
|
\section{Binomial coefficient}
|
|
|
|
\index{binomial coefficient}
|
|
|
|
A \key{binomial coefficient} ${n \choose k}$
|
|
is the number of ways we can choose a subset
|
|
of $k$ elements from a set of $n$ elements.
|
|
For example, ${5 \choose 3}=10$,
|
|
because the set $\{1,2,3,4,5\}$
|
|
has 10 subsets of 3 elements:
|
|
\[ \{1,2,3\}, \{1,2,4\}, \{1,2,5\}, \{1,3,4\}, \{1,3,5\},
|
|
\{1,4,5\}, \{2,3,4\}, \{2,3,5\}, \{2,4,5\}, \{3,4,5\} \]
|
|
|
|
\subsubsection{Formula 1}
|
|
|
|
Binomial coefficients can be
|
|
recursively calculated as follows:
|
|
|
|
\[
|
|
{n \choose k} = {n-1 \choose k-1} + {n-1 \choose k}
|
|
\]
|
|
|
|
The idea is to consider a fixed
|
|
element $x$ in the set.
|
|
If $x$ is included in the subset,
|
|
the remaining task is to choose $k-1$
|
|
elements from $n-1$ elements,
|
|
and otherwise
|
|
the remaining task is to choose $k$ elements from $n-1$ elements.
|
|
|
|
The base cases for the recursion are as follows:
|
|
\[
|
|
{n \choose 0} = {n \choose n} = 1
|
|
\]
|
|
|
|
The reason for this is that there is always
|
|
one way to construct an empty subset,
|
|
or a subset that contains all the elements.
|
|
|
|
\subsubsection{Formula 2}
|
|
|
|
Another way to calculate binomial coefficients is as follows:
|
|
\[
|
|
{n \choose k} = \frac{n!}{k!(n-k)!}.
|
|
\]
|
|
|
|
There are $n!$ permutations for $n$ elements.
|
|
We go through all permutations and in each case
|
|
select the first $k$ elements of the permutation
|
|
to the subset.
|
|
Since the order of the elements in the subset
|
|
and outside the subset doesn't matter,
|
|
the result is divided by $k!$ and $(n-k)!$
|
|
|
|
\subsubsection{Properties}
|
|
|
|
For binomial coefficients,
|
|
\[
|
|
{n \choose k} = {n \choose n-k},
|
|
\]
|
|
because we can either select $k$
|
|
elements to the subset,
|
|
or select $n-k$ elements that
|
|
will be outside the subset.
|
|
|
|
The sum of binomial coefficients is
|
|
\[
|
|
{n \choose 0}+{n \choose 1}+{n \choose 2}+\ldots+{n \choose n}=2^n.
|
|
\]
|
|
|
|
The reason for the name ''binomial coefficient''
|
|
is that
|
|
|
|
\[ (a+b)^n =
|
|
{n \choose 0} a^n b^0 +
|
|
{n \choose 1} a^{n-1} b^1 +
|
|
\ldots +
|
|
{n \choose n-1} a^1 b^{n-1} +
|
|
{n \choose n} a^0 b^n. \]
|
|
|
|
\index{Pascal's triangle}
|
|
|
|
Binomial coefficients also appear in
|
|
\key{Pascal's triangle}
|
|
whose border consists of 1's,
|
|
and each value is the sum of two
|
|
above values:
|
|
\begin{center}
|
|
\begin{tikzpicture}{0.9}
|
|
\node at (0,0) {1};
|
|
\node at (-0.5,-0.5) {1};
|
|
\node at (0.5,-0.5) {1};
|
|
\node at (-1,-1) {1};
|
|
\node at (0,-1) {2};
|
|
\node at (1,-1) {1};
|
|
\node at (-1.5,-1.5) {1};
|
|
\node at (-0.5,-1.5) {3};
|
|
\node at (0.5,-1.5) {3};
|
|
\node at (1.5,-1.5) {1};
|
|
\node at (-2,-2) {1};
|
|
\node at (-1,-2) {4};
|
|
\node at (0,-2) {6};
|
|
\node at (1,-2) {4};
|
|
\node at (2,-2) {1};
|
|
\node at (-2,-2.5) {$\ldots$};
|
|
\node at (-1,-2.5) {$\ldots$};
|
|
\node at (0,-2.5) {$\ldots$};
|
|
\node at (1,-2.5) {$\ldots$};
|
|
\node at (2,-2.5) {$\ldots$};
|
|
\end{tikzpicture}
|
|
\end{center}
|
|
|
|
\subsubsection{Boxes and balls}
|
|
|
|
''Boxes and models'' is a useful model,
|
|
where we count the ways to
|
|
place $k$ balls in $n$ boxes.
|
|
Let's consider three cases:
|
|
|
|
\textit{Case 1}: Each box can contain
|
|
at most one ball.
|
|
For example, when $n=5$ and $k=2$,
|
|
there are 10 solutions:
|
|
|
|
\begin{center}
|
|
\begin{tikzpicture}[scale=0.5]
|
|
\newcommand\lax[3]{
|
|
\path[draw,thick,-] (#1-0.5,#2+0.5) -- (#1-0.5,#2-0.5) --
|
|
(#1+0.5,#2-0.5) -- (#1+0.5,#2+0.5);
|
|
\ifthenelse{\equal{#3}{1}}{\draw[fill=black] (#1,#2-0.3) circle (0.15);}{}
|
|
\ifthenelse{\equal{#3}{2}}{\draw[fill=black] (#1-0.2,#2-0.3) circle (0.15);}{}
|
|
\ifthenelse{\equal{#3}{2}}{\draw[fill=black] (#1+0.2,#2-0.3) circle (0.15);}{}
|
|
}
|
|
\newcommand\laa[7]{
|
|
\lax{#1}{#2}{#3}
|
|
\lax{#1+1.2}{#2}{#4}
|
|
\lax{#1+2.4}{#2}{#5}
|
|
\lax{#1+3.6}{#2}{#6}
|
|
\lax{#1+4.8}{#2}{#7}
|
|
}
|
|
|
|
\laa{0}{0}{1}{1}{0}{0}{0}
|
|
\laa{0}{-2}{1}{0}{1}{0}{0}
|
|
\laa{0}{-4}{1}{0}{0}{1}{0}
|
|
\laa{0}{-6}{1}{0}{0}{0}{1}
|
|
\laa{8}{0}{0}{1}{1}{0}{0}
|
|
\laa{8}{-2}{0}{1}{0}{1}{0}
|
|
\laa{8}{-4}{0}{1}{0}{0}{1}
|
|
\laa{16}{0}{0}{0}{1}{1}{0}
|
|
\laa{16}{-2}{0}{0}{1}{0}{1}
|
|
\laa{16}{-4}{0}{0}{0}{1}{1}
|
|
|
|
\end{tikzpicture}
|
|
\end{center}
|
|
|
|
In this case, the answer is directly the
|
|
binomial coefficient ${n \choose k}$.
|
|
|
|
\textit{Case 2}: A box can contain multiple balls.
|
|
For example, when $n=5$ and $k=2$,
|
|
there are 15 solutions:
|
|
|
|
\begin{center}
|
|
\begin{tikzpicture}[scale=0.5]
|
|
\newcommand\lax[3]{
|
|
\path[draw,thick,-] (#1-0.5,#2+0.5) -- (#1-0.5,#2-0.5) --
|
|
(#1+0.5,#2-0.5) -- (#1+0.5,#2+0.5);
|
|
\ifthenelse{\equal{#3}{1}}{\draw[fill=black] (#1,#2-0.3) circle (0.15);}{}
|
|
\ifthenelse{\equal{#3}{2}}{\draw[fill=black] (#1-0.2,#2-0.3) circle (0.15);}{}
|
|
\ifthenelse{\equal{#3}{2}}{\draw[fill=black] (#1+0.2,#2-0.3) circle (0.15);}{}
|
|
}
|
|
\newcommand\laa[7]{
|
|
\lax{#1}{#2}{#3}
|
|
\lax{#1+1.2}{#2}{#4}
|
|
\lax{#1+2.4}{#2}{#5}
|
|
\lax{#1+3.6}{#2}{#6}
|
|
\lax{#1+4.8}{#2}{#7}
|
|
}
|
|
|
|
\laa{0}{0}{2}{0}{0}{0}{0}
|
|
\laa{0}{-2}{1}{1}{0}{0}{0}
|
|
\laa{0}{-4}{1}{0}{1}{0}{0}
|
|
\laa{0}{-6}{1}{0}{0}{1}{0}
|
|
\laa{0}{-8}{1}{0}{0}{0}{1}
|
|
\laa{8}{0}{0}{2}{0}{0}{0}
|
|
\laa{8}{-2}{0}{1}{1}{0}{0}
|
|
\laa{8}{-4}{0}{1}{0}{1}{0}
|
|
\laa{8}{-6}{0}{1}{0}{0}{1}
|
|
\laa{8}{-8}{0}{0}{2}{0}{0}
|
|
\laa{16}{0}{0}{0}{1}{1}{0}
|
|
\laa{16}{-2}{0}{0}{1}{0}{1}
|
|
\laa{16}{-4}{0}{0}{0}{2}{0}
|
|
\laa{16}{-6}{0}{0}{0}{1}{1}
|
|
\laa{16}{-8}{0}{0}{0}{0}{2}
|
|
|
|
\end{tikzpicture}
|
|
\end{center}
|
|
|
|
This process can be represented as a string
|
|
that consists of symbols
|
|
''o'' and ''$\rightarrow$''.
|
|
Initially, we are standing at the leftmost box.
|
|
The symbol ''o'' means we place a ball
|
|
in the current box, and the symbol
|
|
''$\rightarrow$'' means that we move to
|
|
the next box right.
|
|
|
|
Using this notation, each solution is a string
|
|
that has $k$ times the symbol ''o'' and
|
|
$n-1$ times the symbol ''$\rightarrow$''.
|
|
For example, the upper-right solution
|
|
corresponds to the string
|
|
''$\rightarrow$ $\rightarrow$ o $\rightarrow$ o $\rightarrow$''.
|
|
Thus, the number of solutions is
|
|
${k+n-1 \choose k}$.
|
|
|
|
\textit{Case 3}: Each box may contain at most one ball,
|
|
and in addition, no two adjacent boxes may both contain a ball.
|
|
For example, when $n=5$ and $k=2$,
|
|
there are 6 solutions:
|
|
|
|
|
|
\begin{center}
|
|
\begin{tikzpicture}[scale=0.5]
|
|
\newcommand\lax[3]{
|
|
\path[draw,thick,-] (#1-0.5,#2+0.5) -- (#1-0.5,#2-0.5) --
|
|
(#1+0.5,#2-0.5) -- (#1+0.5,#2+0.5);
|
|
\ifthenelse{\equal{#3}{1}}{\draw[fill=black] (#1,#2-0.3) circle (0.15);}{}
|
|
\ifthenelse{\equal{#3}{2}}{\draw[fill=black] (#1-0.2,#2-0.3) circle (0.15);}{}
|
|
\ifthenelse{\equal{#3}{2}}{\draw[fill=black] (#1+0.2,#2-0.3) circle (0.15);}{}
|
|
}
|
|
\newcommand\laa[7]{
|
|
\lax{#1}{#2}{#3}
|
|
\lax{#1+1.2}{#2}{#4}
|
|
\lax{#1+2.4}{#2}{#5}
|
|
\lax{#1+3.6}{#2}{#6}
|
|
\lax{#1+4.8}{#2}{#7}
|
|
}
|
|
|
|
\laa{0}{0}{1}{0}{1}{0}{0}
|
|
\laa{0}{-2}{1}{0}{0}{1}{0}
|
|
\laa{8}{0}{1}{0}{0}{0}{1}
|
|
\laa{8}{-2}{0}{1}{0}{1}{0}
|
|
\laa{16}{0}{0}{1}{0}{0}{1}
|
|
\laa{16}{-2}{0}{0}{1}{0}{1}
|
|
\end{tikzpicture}
|
|
\end{center}
|
|
|
|
In this case, we can think that
|
|
$k$ balls are initially placed in boxes.
|
|
and between each such box there is an empty box.
|
|
The remaining task is to choose the
|
|
positions for
|
|
$n-k-(k-1)=n-2k+1$ empty boxes.
|
|
There are $k+1$ positions, so as in case 2,
|
|
the number of solutions is
|
|
${n-2k+1+k+1-1 \choose n-2k+1} = {n-k+1 \choose n-2k+1}$.
|
|
|
|
\subsubsection{Multinomial coefficient}
|
|
|
|
\index{multinomial coefficient}
|
|
|
|
A generalization for a binomial coefficient is
|
|
a \key{multinomial coefficient}
|
|
|
|
\[ {n \choose k_1,k_2,\ldots,k_m} = \frac{n!}{k_1! k_2! \cdots k_m!}, \]
|
|
|
|
where $k_1+k_2+\cdots+k_m=n$.
|
|
A multinomial coefficient i the number of ways
|
|
we can divide $n$ elements into subsets
|
|
whose sizes are $k_1,k_2,\ldots,k_m$.
|
|
If $m=2$, the formula
|
|
corresponds to the binomial coefficient formula.
|
|
|
|
\section{Catalan numbers}
|
|
|
|
\index{Catalan number}
|
|
|
|
A \key{Catalan number} $C_n$ is the
|
|
number of valid
|
|
parenthesis expressions that consist of
|
|
$n$ left parentheses and $n$ right parentheses.
|
|
|
|
For example, $C_3=5$, because using three
|
|
left parentheses and three right parentheses,
|
|
we can construct the following parenthesis
|
|
expressions:
|
|
|
|
\begin{itemize}[noitemsep]
|
|
\item \texttt{()()()}
|
|
\item \texttt{(())()}
|
|
\item \texttt{()(())}
|
|
\item \texttt{((()))}
|
|
\item \texttt{(()())}
|
|
\end{itemize}
|
|
|
|
\subsubsection{Parenthesis expressions}
|
|
|
|
\index{parenthesis expression}
|
|
|
|
What is exactly a \emph{valid parenthesis expression}?
|
|
The following rules precisely define all
|
|
valid parenthesis expressions:
|
|
|
|
\begin{itemize}
|
|
\item The expression \texttt{()} is valid.
|
|
\item If a expression $A$ is valid,
|
|
then also the expression
|
|
\texttt{(}$A$\texttt{)} is valid.
|
|
\item If expressions $A$ and $B$ are valid,
|
|
then also the expression $AB$ is valid.
|
|
\end{itemize}
|
|
|
|
Another way to characterize valid
|
|
paranthesis expressions is that if
|
|
we choose any prefix of the expression,
|
|
it has to contain at least as many left
|
|
parentheses as right parentheses.
|
|
In addition, the complete expression has to
|
|
contain an equal number of left and right
|
|
parentheses.
|
|
|
|
\subsubsection{Formula 1}
|
|
|
|
Catalan numbers can be calculated using the formula
|
|
\[ C_n = \sum_{i=0}^{n-1} C_{i} C_{n-i-1}.\]
|
|
|
|
The sum goes through the ways to divide the
|
|
expression into two parts
|
|
such that both parts are valid
|
|
expressions and the first part is as short as possible
|
|
but not empty.
|
|
For any $i$, the first part contains $i+1$ pairs
|
|
of parentheses, and the number of expressions
|
|
is the product of the following values:
|
|
|
|
\begin{itemize}
|
|
\item $C_{i}$: number of ways to construct an expression
|
|
using the parentheses in the first part,
|
|
not counting the outermost parentheses
|
|
\item $C_{n-i-1}$: number of ways to construct an
|
|
expression using the parentheses in the second part
|
|
\end{itemize}
|
|
In addition, the base case is $C_0=1$,
|
|
because we can construct an empty parenthesis
|
|
expression using zero pairs of parentheses.
|
|
|
|
\subsubsection{Formula 2}
|
|
|
|
Catalan numbers can also be calculated
|
|
using binomial coefficients:
|
|
\[ C_n = \frac{1}{n+1} {2n \choose n}\]
|
|
The formula can be explained as follows:
|
|
|
|
There are a total of ${2n \choose n}$ ways
|
|
to construct a (not necessarily valid)
|
|
parenthesis expression that contains $n$ left
|
|
parentheses and $n$ right parentheses.
|
|
Let's calculate the number of such
|
|
expressions that are \emph{not} valid.
|
|
|
|
If a parenthesis expression is not valid,
|
|
it has to contain a prefix where the
|
|
number of right parentheses exceeds the
|
|
number of left parentheses.
|
|
The idea is to reverse each parenthesis
|
|
that belongs to such a prefix.
|
|
For example, the expression
|
|
\texttt{())()(} contains a prefix \texttt{())},
|
|
and after reversing the prefix,
|
|
the expression becomes \texttt{)((()(}.
|
|
|
|
The resulting expression consists of $n+1$
|
|
left parentheses and $n-1$ right parentheses.
|
|
The number of such expressions is ${2n \choose n+1}$
|
|
that equals the number of non-valid
|
|
parenthesis expressions.
|
|
Thus the number of valid parenthesis
|
|
expressions can be calculated using the formula
|
|
\[{2n \choose n}-{2n \choose n+1} = {2n \choose n} - \frac{n}{n+1} {2n \choose n} = \frac{1}{n+1} {2n \choose n}.\]
|
|
|
|
\subsubsection{Counting trees}
|
|
|
|
Catalan numbers are also related to rooted trees:
|
|
|
|
\begin{itemize}
|
|
\item there are $C_n$ binary trees of $n$ nodes
|
|
\item there are $C_{n-1}$ rooted trees of $n$ nodes
|
|
\end{itemize}
|
|
\noindent
|
|
For example, for $C_3=5$, the binary trees are
|
|
|
|
\begin{center}
|
|
\begin{tikzpicture}[scale=0.7]
|
|
\path[draw,thick,-] (0,0) -- (-1,-1);
|
|
\path[draw,thick,-] (0,0) -- (1,-1);
|
|
\draw[fill=white] (0,0) circle (0.3);
|
|
\draw[fill=white] (-1,-1) circle (0.3);
|
|
\draw[fill=white] (1,-1) circle (0.3);
|
|
|
|
\path[draw,thick,-] (4,0) -- (4-0.75,-1) -- (4-1.5,-2);
|
|
\draw[fill=white] (4,0) circle (0.3);
|
|
\draw[fill=white] (4-0.75,-1) circle (0.3);
|
|
\draw[fill=white] (4-1.5,-2) circle (0.3);
|
|
|
|
\path[draw,thick,-] (6.5,0) -- (6.5-0.75,-1) -- (6.5-0,-2);
|
|
\draw[fill=white] (6.5,0) circle (0.3);
|
|
\draw[fill=white] (6.5-0.75,-1) circle (0.3);
|
|
\draw[fill=white] (6.5-0,-2) circle (0.3);
|
|
|
|
\path[draw,thick,-] (9,0) -- (9+0.75,-1) -- (9-0,-2);
|
|
\draw[fill=white] (9,0) circle (0.3);
|
|
\draw[fill=white] (9+0.75,-1) circle (0.3);
|
|
\draw[fill=white] (9-0,-2) circle (0.3);
|
|
|
|
\path[draw,thick,-] (11.5,0) -- (11.5+0.75,-1) -- (11.5+1.5,-2);
|
|
\draw[fill=white] (11.5,0) circle (0.3);
|
|
\draw[fill=white] (11.5+0.75,-1) circle (0.3);
|
|
\draw[fill=white] (11.5+1.5,-2) circle (0.3);
|
|
\end{tikzpicture}
|
|
\end{center}
|
|
and the rooted trees are
|
|
\begin{center}
|
|
\begin{tikzpicture}[scale=0.7]
|
|
\path[draw,thick,-] (0,0) -- (-1,-1);
|
|
\path[draw,thick,-] (0,0) -- (0,-1);
|
|
\path[draw,thick,-] (0,0) -- (1,-1);
|
|
\draw[fill=white] (0,0) circle (0.3);
|
|
\draw[fill=white] (-1,-1) circle (0.3);
|
|
\draw[fill=white] (0,-1) circle (0.3);
|
|
\draw[fill=white] (1,-1) circle (0.3);
|
|
|
|
\path[draw,thick,-] (3,0) -- (3,-1) -- (3,-2) -- (3,-3);
|
|
\draw[fill=white] (3,0) circle (0.3);
|
|
\draw[fill=white] (3,-1) circle (0.3);
|
|
\draw[fill=white] (3,-2) circle (0.3);
|
|
\draw[fill=white] (3,-3) circle (0.3);
|
|
|
|
\path[draw,thick,-] (6+0,0) -- (6-1,-1);
|
|
\path[draw,thick,-] (6+0,0) -- (6+1,-1) -- (6+1,-2);
|
|
\draw[fill=white] (6+0,0) circle (0.3);
|
|
\draw[fill=white] (6-1,-1) circle (0.3);
|
|
\draw[fill=white] (6+1,-1) circle (0.3);
|
|
\draw[fill=white] (6+1,-2) circle (0.3);
|
|
|
|
\path[draw,thick,-] (9+0,0) -- (9+1,-1);
|
|
\path[draw,thick,-] (9+0,0) -- (9-1,-1) -- (9-1,-2);
|
|
\draw[fill=white] (9+0,0) circle (0.3);
|
|
\draw[fill=white] (9+1,-1) circle (0.3);
|
|
\draw[fill=white] (9-1,-1) circle (0.3);
|
|
\draw[fill=white] (9-1,-2) circle (0.3);
|
|
|
|
\path[draw,thick,-] (12+0,0) -- (12+0,-1) -- (12-1,-2);
|
|
\path[draw,thick,-] (12+0,0) -- (12+0,-1) -- (12+1,-2);
|
|
\draw[fill=white] (12+0,0) circle (0.3);
|
|
\draw[fill=white] (12+0,-1) circle (0.3);
|
|
\draw[fill=white] (12-1,-2) circle (0.3);
|
|
\draw[fill=white] (12+1,-2) circle (0.3);
|
|
|
|
\end{tikzpicture}
|
|
\end{center}
|
|
|
|
|
|
\section{Inkluusio-ekskluusio}
|
|
|
|
\index{inkluusio-ekskluusio}
|
|
|
|
\key{Inkluusio-ekskluusio}
|
|
on tekniikka, jonka avulla pystyy laskemaan
|
|
joukkojen yhdisteen koon leikkausten
|
|
kokojen perusteella ja päinvastoin.
|
|
Yksinkertainen esimerkki periaatteesta on kaava
|
|
\[ |A \cup B| = |A| + |B| - |A \cap B|,\]
|
|
jossa $A$ ja $B$ ovat joukkoja ja $|X|$
|
|
tarkoittaa joukon $X$ kokoa.
|
|
Seuraava kuva havainnollistaa kaavaa,
|
|
kun joukot ovat tason ympyröitä:
|
|
|
|
\begin{center}
|
|
\begin{tikzpicture}[scale=0.8]
|
|
|
|
\draw (0,0) circle (1.5);
|
|
\draw (1.5,0) circle (1.5);
|
|
|
|
\node at (-0.75,0) {\small $A$};
|
|
\node at (2.25,0) {\small $B$};
|
|
\node at (0.75,0) {\small $A \cap B$};
|
|
|
|
\end{tikzpicture}
|
|
\end{center}
|
|
|
|
Tavoitteena on laskea, kuinka suuri on yhdiste $A \cup B$
|
|
eli alue, joka on toisen tai kummankin ympyrän sisällä.
|
|
Kuvan mukaisesti yhdisteen $A \cup B$ koko
|
|
saadaan laskemalla ensin yhteen ympyröiden $A$ ja $B$ koot
|
|
ja vähentämällä siitä sitten leikkauksen $A \cap B$ koko.
|
|
|
|
Samaa ideaa voi soveltaa, kun joukkoja on enemmän.
|
|
Kolmen joukon tapauksessa kaavasta tulee
|
|
\[ |A \cup B \cup C| = |A| + |B| + |C| - |A \cap B| - |A \cap C| - |B \cap C| + |A \cap B \cap C| \]
|
|
ja vastaava kuva on
|
|
|
|
\begin{center}
|
|
\begin{tikzpicture}[scale=0.8]
|
|
|
|
\draw (0,0) circle (1.75);
|
|
\draw (2,0) circle (1.75);
|
|
\draw (1,1.5) circle (1.75);
|
|
|
|
\node at (-0.75,-0.25) {\small $A$};
|
|
\node at (2.75,-0.25) {\small $B$};
|
|
\node at (1,2.5) {\small $C$};
|
|
\node at (1,-0.5) {\small $A \cap B$};
|
|
\node at (0,1.25) {\small $A \cap C$};
|
|
\node at (2,1.25) {\small $B \cap C$};
|
|
\node at (1,0.5) {\scriptsize $A \cap B \cap C$};
|
|
|
|
\end{tikzpicture}
|
|
\end{center}
|
|
|
|
Yleisessä tapauksessa yhdisteen $X_1 \cup X_2 \cup \cdots \cup X_n$
|
|
koon saa laskettua käymällä läpi kaikki tavat muodostaa
|
|
leikkaus joukoista $X_1,X_2,\ldots,X_n$.
|
|
Parittoman määrän joukkoja sisältävät leikkaukset
|
|
lasketaan mukaan positiivisina ja
|
|
parillisen määrän negatiivisina.
|
|
|
|
Huomaa, että vastaavat kaavat toimivat myös käänteisesti
|
|
leikkauksen koon laskemiseen yhdisteiden kokojen perusteella.
|
|
Esimerkiksi
|
|
\[ |A \cap B| = |A| + |B| - |A \cup B|\]
|
|
ja
|
|
\[ |A \cap B \cap C| = |A| + |B| + |C| - |A \cup B| - |A \cup C| - |B \cup C| + |A \cup B \cup C| .\]
|
|
|
|
\subsubsection{Epäjärjestykset}
|
|
|
|
\index{epxjxrjestys@epäjärjestys}
|
|
|
|
Lasketaan esimerkkinä,
|
|
montako tapaa on muodostaa luvuista
|
|
$(1,2,\ldots,n)$ \key{epäjärjestys}
|
|
eli permutaatio,
|
|
jossa mikään luku ei ole alkuperäisellä paikallaan.
|
|
Esimerkiksi jos $n=3$, niin epäjärjestyksiä on kaksi: $(2,3,1)$ ja $(3,1,2)$.
|
|
|
|
Yksi tapa lähestyä tehtävää on käyttää inkluusio-ekskluusiota.
|
|
Olkoon joukko $X_k$ niiden permutaatioiden joukko,
|
|
jossa kohdassa $k$ on luku $k$.
|
|
Esimerkiksi jos $n=3$, niin joukot ovat seuraavat:
|
|
\[
|
|
\begin{array}{lcl}
|
|
X_1 & = & \{(1,2,3),(1,3,2)\} \\
|
|
X_2 & = & \{(1,2,3),(3,2,1)\} \\
|
|
X_3 & = & \{(1,2,3),(2,1,3)\} \\
|
|
\end{array}
|
|
\]
|
|
Näitä joukkoja käyttäen epäjärjestysten määrä on
|
|
\[ n! - |X_1 \cup X_2 \cup \cdots \cup X_n|, \]
|
|
eli
|
|
riittää laskea joukkojen yhdisteen koko.
|
|
Tämä palautuu inkluusio-eks\-kluu\-sion avulla
|
|
joukkojen leikkausten kokojen laskemiseen,
|
|
mikä onnistuu tehokkaasti.
|
|
Esimerkiksi kun $n=3$, joukon $|X_1 \cup X_2 \cup X_3|$ koko on
|
|
\[
|
|
\begin{array}{lcl}
|
|
& & |X_1| + |X_2| + |X_3| - |X_1 \cap X_2| - |X_1 \cap X_3| - |X_2 \cap X_3| + |X_1 \cap X_2 \cap X_3| \\
|
|
& = & 2+2+2-1-1-1+1 \\
|
|
& = & 4, \\
|
|
\end{array}
|
|
\]
|
|
joten ratkaisujen määrä on $3!-4=2$.
|
|
|
|
Osoittautuu, että tehtävän voi ratkaista myös toisella
|
|
tavalla käyttämättä inkluusio-ekskluusiota.
|
|
Merkitään $f(n)$:llä jonon $(1,2,\ldots,n)$ epäjärjestysten määrää,
|
|
jolloin seuraava rekursio pätee:
|
|
|
|
\begin{equation*}
|
|
f(n) = \begin{cases}
|
|
0 & n = 1\\
|
|
1 & n = 2\\
|
|
(n-1)(f(n-2) + f(n-1)) & n>2 \\
|
|
\end{cases}
|
|
\end{equation*}
|
|
|
|
Kaavan voi perustella käymällä läpi tapaukset,
|
|
miten luku 1 muuttuu epäjärjestyksessä.
|
|
On $n-1$ tapaa valita jokin luku $x$ luvun 1 tilalle.
|
|
Jokaisessa tällaisessa valinnassa on kaksi vaihtoehtoa:
|
|
|
|
\textit{Vaihtoehto 1:} Luvun $x$ tilalle valitaan luku 1.
|
|
Tällöin jää $n-2$ lukua, joille tulee muodostaa epäjärjestys.
|
|
|
|
\textit{Vaihtoehto 2:} Luvun $x$ tilalle ei valita lukua 1.
|
|
Tällöin jää $n-1$ lukua, joille tulee muodostaa epäjärjestys,
|
|
koska luvun $x$ tilalle ei saa valita lukua 1
|
|
ja kaikki muut luvut tulee saattaa epäjärjestykseen.
|
|
|
|
\section{Burnsiden lemma}
|
|
|
|
\index{Burnsiden lemma@Burnsiden lemma}
|
|
|
|
\key{Burnsiden lemma} laskee yhdistelmien määrän niin,
|
|
että symmetrisistä yhdistelmistä lasketaan
|
|
mukaan vain yksi edustaja.
|
|
Burnsiden lemman mukaan yhdistelmien määrä on
|
|
\[\sum_{k=1}^n \frac{c(k)}{n},\]
|
|
missä yhdistelmän asentoa voi muuttaa $n$ tavalla
|
|
ja $c(k)$ on niiden yhdistelmien määrä,
|
|
jotka pysyvät ennallaan, kun asentoa
|
|
muutetaan tavalla $k$.
|
|
|
|
Lasketaan esimerkkinä, montako
|
|
erilaista tapaa on
|
|
muodostaa $n$ helmen helminauha,
|
|
kun kunkin helmen värin tulee olla
|
|
väliltä $1,2,\ldots,m$.
|
|
Kaksi helminauhaa ovat symmetriset,
|
|
jos ne voi saada näyttämään samalta pyörittämällä.
|
|
Esimerkiksi helminauhan
|
|
\begin{center}
|
|
\begin{tikzpicture}[scale=0.7]
|
|
\draw[fill=white] (0,0) circle (1);
|
|
\draw[fill=red] (0,1) circle (0.3);
|
|
\draw[fill=blue] (1,0) circle (0.3);
|
|
\draw[fill=red] (0,-1) circle (0.3);
|
|
\draw[fill=green] (-1,0) circle (0.3);
|
|
\end{tikzpicture}
|
|
\end{center}
|
|
kanssa symmetriset helminauhat ovat seuraavat:
|
|
\begin{center}
|
|
\begin{tikzpicture}[scale=0.7]
|
|
\draw[fill=white] (0,0) circle (1);
|
|
\draw[fill=red] (0,1) circle (0.3);
|
|
\draw[fill=blue] (1,0) circle (0.3);
|
|
\draw[fill=red] (0,-1) circle (0.3);
|
|
\draw[fill=green] (-1,0) circle (0.3);
|
|
|
|
\draw[fill=white] (4,0) circle (1);
|
|
\draw[fill=green] (4+0,1) circle (0.3);
|
|
\draw[fill=red] (4+1,0) circle (0.3);
|
|
\draw[fill=blue] (4+0,-1) circle (0.3);
|
|
\draw[fill=red] (4+-1,0) circle (0.3);
|
|
|
|
\draw[fill=white] (8,0) circle (1);
|
|
\draw[fill=red] (8+0,1) circle (0.3);
|
|
\draw[fill=green] (8+1,0) circle (0.3);
|
|
\draw[fill=red] (8+0,-1) circle (0.3);
|
|
\draw[fill=blue] (8+-1,0) circle (0.3);
|
|
|
|
\draw[fill=white] (12,0) circle (1);
|
|
\draw[fill=blue] (12+0,1) circle (0.3);
|
|
\draw[fill=red] (12+1,0) circle (0.3);
|
|
\draw[fill=green] (12+0,-1) circle (0.3);
|
|
\draw[fill=red] (12+-1,0) circle (0.3);
|
|
\end{tikzpicture}
|
|
\end{center}
|
|
Tapoja muuttaa asentoa on $n$,
|
|
koska helminauhaa voi pyörittää $0,1,\ldots,n-1$
|
|
askelta myötäpäivään.
|
|
Jos helminauhaa pyörittää 0 askelta,
|
|
kaikki $m^n$ väritystä säilyvät ennallaan.
|
|
Jos taas helminauhaa pyörittää 1 askeleen,
|
|
vain $m$ yksiväristä helminauhaa säilyy ennallaan.
|
|
|
|
Yleisemmin kun helminauhaa pyörittää $k$ askelta,
|
|
ennallaan säilyvien yhdistelmien määrä on
|
|
\[m^{\textrm{syt}(k,n)},\]
|
|
missä $\textrm{syt}(k,n)$ on lukujen $k$ ja $n$
|
|
suurin yhteinen tekijä.
|
|
Tämä johtuu siitä, että $\textrm{syt}(k,n)$-kokoiset
|
|
pätkät helmiä siirtyvät toistensa paikoille
|
|
$k$ askelta eteenpäin.
|
|
Niinpä helminauhojen määrä on
|
|
Burnsiden lemman mukaan
|
|
\[\sum_{i=0}^{n-1} \frac{m^{\textrm{syt}(i,n)}}{n}. \]
|
|
Esimerkiksi kun helminauhan pituus on 4
|
|
ja värejä on 3, helminauhoja on
|
|
\[\frac{3^4+3+3^2+3}{4} = 24. \]
|
|
|
|
\section{Cayleyn kaava}
|
|
|
|
\index{Cayleyn kaava@Cayleyn kaava}
|
|
|
|
\key{Cayleyn kaavan} mukaan $n$ solmusta voi
|
|
muodostaa $n^{n-2}$ numeroitua puuta.
|
|
Puun solmut on numeroitu $1,2,\ldots,n$,
|
|
ja kaksi puuta ovat erilaiset,
|
|
jos niiden rakenne on erilainen
|
|
tai niissä on eri numerointi.
|
|
|
|
\begin{samepage}
|
|
\noindent
|
|
Esimerkiksi kun $n=4$, numeroitujen puiden määrä on $4^{4-2}=16$:
|
|
|
|
\begin{center}
|
|
\begin{tikzpicture}[scale=0.8]
|
|
\footnotesize
|
|
|
|
\newcommand\puua[6]{
|
|
\path[draw,thick,-] (#1,#2) -- (#1-1.25,#2-1.5);
|
|
\path[draw,thick,-] (#1,#2) -- (#1,#2-1.5);
|
|
\path[draw,thick,-] (#1,#2) -- (#1+1.25,#2-1.5);
|
|
\node[draw, circle, fill=white] at (#1,#2) {#3};
|
|
\node[draw, circle, fill=white] at (#1-1.25,#2-1.5) {#4};
|
|
\node[draw, circle, fill=white] at (#1,#2-1.5) {#5};
|
|
\node[draw, circle, fill=white] at (#1+1.25,#2-1.5) {#6};
|
|
}
|
|
\newcommand\puub[6]{
|
|
\path[draw,thick,-] (#1,#2) -- (#1+1,#2);
|
|
\path[draw,thick,-] (#1+1,#2) -- (#1+2,#2);
|
|
\path[draw,thick,-] (#1+2,#2) -- (#1+3,#2);
|
|
\node[draw, circle, fill=white] at (#1,#2) {#3};
|
|
\node[draw, circle, fill=white] at (#1+1,#2) {#4};
|
|
\node[draw, circle, fill=white] at (#1+2,#2) {#5};
|
|
\node[draw, circle, fill=white] at (#1+3,#2) {#6};
|
|
}
|
|
|
|
\puua{0}{0}{1}{2}{3}{4}
|
|
\puua{4}{0}{2}{1}{3}{4}
|
|
\puua{8}{0}{3}{1}{2}{4}
|
|
\puua{12}{0}{4}{1}{2}{3}
|
|
|
|
\puub{0}{-3}{1}{2}{3}{4}
|
|
\puub{4.5}{-3}{1}{2}{4}{3}
|
|
\puub{9}{-3}{1}{3}{2}{4}
|
|
\puub{0}{-4.5}{1}{3}{4}{2}
|
|
\puub{4.5}{-4.5}{1}{4}{2}{3}
|
|
\puub{9}{-4.5}{1}{4}{3}{2}
|
|
\puub{0}{-6}{2}{1}{3}{4}
|
|
\puub{4.5}{-6}{2}{1}{4}{3}
|
|
\puub{9}{-6}{2}{3}{1}{4}
|
|
\puub{0}{-7.5}{2}{4}{1}{3}
|
|
\puub{4.5}{-7.5}{3}{1}{2}{4}
|
|
\puub{9}{-7.5}{3}{2}{1}{4}
|
|
\end{tikzpicture}
|
|
\end{center}
|
|
\end{samepage}
|
|
|
|
Seuraavaksi näemme, miten Cayleyn kaavan
|
|
voi perustella samastamalla numeroidut puut
|
|
Prüfer-koodeihin.
|
|
|
|
\subsubsection{Prüfer-koodi}
|
|
|
|
\index{Prüfer-koodi}
|
|
|
|
\key{Prüfer-koodi} on $n-2$ luvun jono,
|
|
joka kuvaa numeroidun puun rakenteen.
|
|
Koodi muodostuu poistamalla puusta
|
|
joka askeleella lehden, jonka numero on pienin,
|
|
ja lisäämällä lehden vieressä olevan solmun
|
|
numeron koodiin.
|
|
|
|
Esimerkiksi puun
|
|
\begin{center}
|
|
\begin{tikzpicture}[scale=0.9]
|
|
\node[draw, circle] (1) at (2,3) {$1$};
|
|
\node[draw, circle] (2) at (4,3) {$2$};
|
|
\node[draw, circle] (3) at (2,1) {$3$};
|
|
\node[draw, circle] (4) at (4,1) {$4$};
|
|
\node[draw, circle] (5) at (5.5,2) {$5$};
|
|
|
|
%\path[draw,thick,-] (1) -- (2);
|
|
%\path[draw,thick,-] (1) -- (3);
|
|
\path[draw,thick,-] (1) -- (4);
|
|
\path[draw,thick,-] (3) -- (4);
|
|
\path[draw,thick,-] (2) -- (4);
|
|
\path[draw,thick,-] (2) -- (5);
|
|
%\path[draw,thick,-] (4) -- (5);
|
|
\end{tikzpicture}
|
|
\end{center}
|
|
Prüfer-koodi on $[4,4,2]$,
|
|
koska puusta poistetaan ensin solmu 1,
|
|
sitten solmu 3 ja lopuksi solmu 5.
|
|
|
|
Jokaiselle puulle voidaan laskea
|
|
Prüfer-koodi, minkä lisäksi
|
|
Prüfer-koodista pystyy palauttamaan
|
|
yksikäsitteisesti alkuperäisen puun.
|
|
Niinpä numeroituja puita on yhtä monta
|
|
kuin Prüfer-koodeja eli $n^{n-2}$.
|
|
|