# Mixing

2010 Mathematics Subject Classification: Primary: 37A25 [MSN][ZBL]

A property of a dynamical system (a cascade $\{ S ^ {n} \}$ or a flow (continuous-time dynamical system) $\{ S _ {t} \}$) having a finite invariant measure $\mu$, in which for any two measurable subsets $A$ and $B$ of the phase space $W$, the measure

$$\mu (( S ^ {n} ) ^ {-} 1 A \cap B),$$

or, respectively,

$$\mu (( S _ {t} ) ^ {-} 1 A \cap B),$$

tends to

$$\frac{\mu ( A) \mu ( B) }{\mu ( W) }$$

as $n \rightarrow \infty$, or, respectively, as $t \rightarrow \infty$. If the transformations $S$ and $S _ {t}$ are invertible, then in the definition of mixing one may replace the pre-images of the original set $A$ with respect to these transformations by the direct images $S ^ {n} A$ and $S _ {t} A$, which are easier to visualize. If a system has the mixing property, one also says that the system is mixing, while in the case of a mixing cascade $\{ S ^ {n} \}$, one says that the endomorphism $S$ generating it in the measure space $( W, \mu )$ also is mixing (has the property of mixing).

In ergodic theory, properties related to mixing are considered: multiple mixing and weak mixing (see [H]; in the old literature the latter is often called mixing in the wide sense or simply mixing, while mixing was called mixing in the strong sense). A property intermediate between mixing and weak mixing has also been discussed [FW]. All these properties are stronger than ergodicity.

There is an analogue of mixing for systems having an infinite invariant measure [KS].

#### References

 [H] P.R. Halmos, "Lectures on ergodic theory" , Math. Soc. Japan (1956) MR0097489 Zbl 0073.09302 [FW] H. Furstenberg, B. Weiss, "The finite multipliers of infinite ergodic transformations" N.G. Markley (ed.) J.C. Martin (ed.) W. Perrizo (ed.) , The Structure of Attractors in Dynamical Systems , Lect. notes in math. , 668 , Springer (1978) pp. 127–132 MR0518553 Zbl 0385.28009 [KS] U. Krengel, L. Sucheston, "On mixing in infinite measure spaces" Z. Wahrscheinlichkeitstheor. Verw. Geb. , 13 : 2 (1969) pp. 150–164 MR0254215 Zbl 0176.33804

For a cascade $\{ S ^ {n} \}$ on a finite measure space $( W , \mu )$ the notion of weak mixing as defined above is equivalent to the property that the cascade generated by $S \times S$ on the measure space $( W \times W , \mu \otimes \mu )$, where $\mu \otimes \mu$ denotes the product measure, is ergodic (cf. Ergodicity; Metric transitivity). See [H].
For topological dynamical systems the notions of strong and weak mixing have been defined as well [F]. A flow on a topological space $W$ is said to be topologically weakly mixing whenever the flow $\{ S _ {t} \times S _ {t} \}$ on $W \times W$( with the usual product topology) is topologically ergodic; equivalently: whenever for every choice of four non-empty open subsets $U _ {i} , V _ {i}$( $i = 1 , 2$) of $W$ where exists a $t$ such that $S _ {t} U _ {i} \cap V _ {i} \neq \emptyset$ for $i = 1 , 2$. On compact spaces the weakly mixing minimal flows are the minimal flows that have no non-trivial equicontinuous factors; see [A], p. 133. A flow $\{ S _ {t} \}$ on a space $W$ is said to be topologically strongly mixing whenever for every two non-empty open subsets $U$ and $V$ of $W$ there exists a value $t _ {0}$ such that $S _ {t} U \cap V \neq \emptyset$ for all $| t | \geq t _ {0}$. For example, the geodesic flow on a complete two-dimensional Riemannian manifold of constant negative curvature is topologically strongly mixing; see [GH], 13.49. For cascades, the definitions are analogous.