Paper 4, Section II, E

Dynamical Systems
Part II, 2009

Let I,JI, J be closed bounded intervals in R\mathbb{R}, and let F:RRF: \mathbb{R} \rightarrow \mathbb{R} be a continuous map.

Explain what is meant by the statement that ' IFI F-covers JJ ' (written IJ)I \rightarrow J). For a collection of intervals I0,,IkI_{0}, \ldots, I_{k} define the associated directed graph Γ\Gamma and transition matrix AA. Derive an expression for the number of (not necessarily least) period- nn points of FF in terms of AA.

Let FF have a 5 -cycle

x0<x1<x2<x3<x4x_{0}<x_{1}<x_{2}<x_{3}<x_{4}

such that xi+1=F(xi)x_{i+1}=F\left(x_{i}\right) for i=0,,4i=0, \ldots, 4 where indices are taken modulo 5 . Write down the directed graph Γ\Gamma and transition matrix AA for the FF-covering relations between the intervals [xi,xi+1]\left[x_{i}, x_{i+1}\right]. Compute the number of nn-cycles which are guaranteed to exist for FF, for each integer 1n41 \leqslant n \leqslant 4, and the intervals the points move between.

Explain carefully whether or not FF is guaranteed to have a horseshoe. Must FF be chaotic? Could FF be a unimodal map? Justify your answers.

Similarly, a continuous map G:RRG: \mathbb{R} \rightarrow \mathbb{R} has a 5 -cycle

x3<x1<x0<x2<x4x_{3}<x_{1}<x_{0}<x_{2}<x_{4}

For what integer values of n,1n4n, 1 \leqslant n \leqslant 4, is GG guaranteed to have an nn-cycle?

Is GG guaranteed to have a horseshoe? Must GG be chaotic? Justify your answers.