PlanetMath (more info)
 Math for the people, by the people. Sponsor PlanetMath
Encyclopedia | Requests | Forums | Docs | Wiki | Random | RSS  
Login
create new user
name:
pass:
forget your password?
Main Menu
Owner confidence rating: Very high Entry average rating: No information on entry rating
[parent] every permutation has a cycle decomposition (Proof)

In this entry, we shall show that every permutation of a finite set can be factored into a product of disjoint cycles. To accomplish this, we shall proceed in two steps.

We begin by showing that, if $f$ is a non-trivial permutation of a set $\{ x_i \mid 1 \le i \le n \}$ , then there exists a cycle $(y_1, \ldots y_m)$ where$$ \{ y_i \mid 1 \le i \le m \} \subseteq \{ x_i \mid 1 \le i \le n \}$$ and a permutation $g$ of $\{ x_i \mid 1 \le i \le n \}$ such that $f = (y_1, \ldots y_m) \circ g$ and $g(y_i) = y_i$ .

Since the permutation is not trivial, there exists $z$ such that $f(z) \neq z$ . Define a sequence inductively as follows:

$\displaystyle z_1$ $\displaystyle = z$    
$\displaystyle z_{k+1}$ $\displaystyle = f(z_k)$    

Note that we cannot have $z_{k+1} = z_k$ for any $k$ . This follows from a simple induction argument. By definition $f(z_1) = f(z) \neq z = z_1$ . Suppose that $f(z_k) \neq z_k$ but that $f(z_{k+1}) = z_{k+1}$ . By definition, $f(z_k) = z_{k+1}$ . Since $f$ is a permutation, $f(z_k) = z_{k+1}$ and $f(z_{k+1}) = z_{k+1}$ imply that $z_k = z_{k+1}$ , so $z_k = f(z_k)$ , which contradicts a hypothesis. Hence, if $f(z_k) \neq z_k$ , then $f(z_{k+1}) \neq z_{k+1}$ so, by induction, $f(z_k) \neq z_k$ for all $k$ .

By the pigeonhole principle, there must exist $p$ and $q$ such that $p < q$ but $f(z_p) = f(z_q)$ . Let $m$ be the least integer such that $f(z_p) = f(z_{p+m})$ but $f(z_p) \neq f(s_{p+k})$ when $k < m$ . Set $y_k = z_{k+p}$ . Then we have that $(y_1, \ldots y_m)$ is a cycle.

Since $f$ is a permutation and $\{ y_i \mid 1 \le i \le m \}$ is closed under $f$ , it follows that$$ \{ x_i \mid 1 \le i \le n \} \setminus \{ y_i \mid 1 \le i \le m \}$$ is also closed under $f$ . Define $g$ as follows:

\begin{displaymath} g(z) = \begin{cases} z & z \in \{ y_i \mid 1 \le i \le m \} ... ...e i \le n \} \setminus \{ y_i \mid 1 \le i \le m \} \end{cases}\end{displaymath}
Then it is easily verified that $f = (y_1, \ldots y_m) \circ g$ .

We are now in a position to finish the proof that every permutation can be decomposed into cycles. Trivially, a permutation of a set with one element can be decomposed into cycles because the only permutation of a set with one element is the identity permutation, which requires no cycles to decompose. Next, suppose that any set with less than $n$ elements can be decomposed into cycles. Let $f$ be a permutation on a set with $n$ elements. Then, by what we have shown, $f$ can be written as the product of a cycle and a permutation $g$ which fixes the elements of the cycle. The restriction of $g$ to those elements $z$ such that $g(z) \neq z$ is a permutation on less than $n$ elements and hence, by our supposition, can be decomposed into cycles. Thus, $f$ can also be decomposed into cycles. By induction, we conclude that any permutation of a finite set can be decomposed into cycles.




"every permutation has a cycle decomposition" is owned by rspuzio.
(view preamble | get metadata)

View style:


This object's parent.
Log in to rate this entry.
(view current ratings)

Cross-references: restriction, fixes, identity, proof, closed under, integer, pigeonhole principle, hypothesis, imply, argument, induction, simple, sequence, cycles, disjoint, product, finite set, permutation
There is 1 reference to this entry.

This is version 7 of every permutation has a cycle decomposition, born on 2007-03-07, modified 2007-03-07.
Object id is 9045, canonical name is EveryPermutationHasACycleDecomposition.
Accessed 2152 times total.

Classification:
AMS MSC20F55 (Group theory and generalizations :: Special aspects of infinite or finite groups :: Reflection and Coxeter groups)
 05A05 (Combinatorics :: Enumerative combinatorics :: Combinatorial choice problems )
 03-00 (Mathematical logic and foundations :: General reference works )

Pending Errata and Addenda
None.
[ View all 1 ]
Discussion
Style: Expand: Order:
forum policy

No messages.

Interact
post | correct | update request | add example | add (any)