# Limit set

In mathematics, especially in the study of dynamical systems, a limit set is the state a dynamical system reaches after an infinite amount of time has passed, by either going forward or backwards in time. Limit sets are important because they can be used to understand the long term behavior of a dynamical system.

## Types

In general limits sets can be very complicated as in the case of strange attractors, but for 2-dimensional dynamical systems the Poincaré–Bendixson theorem provides a simple characterization of all possible limit sets as a union of fixed points and periodic orbits.

## Definition for iterated functions

Let ${\displaystyle X}$ be a metric space, and let ${\displaystyle f:X\rightarrow X}$ be a continuous function. The ${\displaystyle \omega }$-limit set of ${\displaystyle x\in X}$, denoted by ${\displaystyle \omega (x,f)}$, is the set of cluster points of the forward orbit ${\displaystyle \{f^{n}(x)\}_{n\in \mathbb {N} }}$ of the iterated function ${\displaystyle f}$. Hence, ${\displaystyle y\in \omega (x,f)}$ if and only if there is a strictly increasing sequence of natural numbers ${\displaystyle \{n_{k}\}_{k\in \mathbb {N} }}$ such that ${\displaystyle f^{n_{k}}(x)\rightarrow y}$ as ${\displaystyle k\rightarrow \infty }$. Another way to express this is

${\displaystyle \omega (x,f)=\bigcap _{n\in \mathbb {N} }{\overline {\{f^{k}(x):k>n\}}},}$

where ${\displaystyle {\overline {S}}}$ denotes the closure of set ${\displaystyle S}$. The closure is here needed, since we have not assumed that the underlying metric space of interest to be a complete metric space. The points in the limit set are non-wandering (but may not be recurrent points). This may also be formulated as the outer limit (limsup) of a sequence of sets, such that

${\displaystyle \omega (x,f)=\bigcap _{n=1}^{\infty }{\overline {\bigcup _{k=n}^{\infty }\{f^{k}(x)\}}}.}$

If ${\displaystyle f}$ is a homeomorphism (that is, a bicontinuous bijection), then the ${\displaystyle \alpha }$-limit set is defined in a similar fashion, but for the backward orbit; i.e. ${\displaystyle \alpha (x,f)=\omega (x,f^{-1})}$.

Both sets are ${\displaystyle f}$-invariant, and if ${\displaystyle X}$ is compact, they are compact and nonempty.

## Definition for flows

Given a real dynamical system (T, X, φ) with flow ${\displaystyle \varphi :\mathbb {R} \times X\to X}$, a point x and an orbit γ through x, we call a point y an ω-limit point of x if there exists a sequence ${\displaystyle (t_{n})_{n\in \mathbb {N} }}$ in R so that

${\displaystyle \lim _{n\to \infty }t_{n}=\infty }$
${\displaystyle \lim _{n\to \infty }\varphi (t_{n},x)=y}$.

Analogously we call y an α-limit point of x if there exists a sequence ${\displaystyle (t_{n})_{n\in \mathbb {N} }}$ in R so that

${\displaystyle \lim _{n\to \infty }t_{n}=-\infty }$
${\displaystyle \lim _{n\to \infty }\varphi (t_{n},x)=y}$.

The set of all ω-limit points (α-limit points) for a given orbit γ is called ω-limit set (α-limit set) for γ and denoted limω γ (limα γ).

If the ω-limit set (α-limit set) is disjoint from the orbit γ, that is limω γ ∩ γ = ∅ (limα γ ∩ γ = ∅), we call limω γ (limα γ) a ω-limit cycle (α-limit cycle).

Alternatively the limit sets can be defined as

${\displaystyle \lim _{\omega }\gamma :=\bigcap _{s\in \mathbb {R} }{\overline {\{\varphi (x,t):t>s\}}}}$

and

${\displaystyle \lim _{\alpha }\gamma :=\bigcap _{s\in \mathbb {R} }{\overline {\{\varphi (x,t):t

### Examples

• For any periodic orbit γ of a dynamical system, limω γ = limα γ = γ
• For any fixed point ${\displaystyle x_{0}}$ of a dynamical system, limω ${\displaystyle x_{0}}$ = limα ${\displaystyle x_{0}}$ = ${\displaystyle x_{0}}$

### Properties

• limω γ and limα γ are closed
• if X is compact then limω γ and limα γ are nonempty, compact and connected
• limω γ and limα γ are φ-invariant, that is φ(R × limω γ) = limω γ and φ(R × limα γ) = limα γ