Let $(X,d)$ is a metric space and $X$ has no isolated points, $T:X\rightarrow X$ is a continuous self-map.

Def1. $T$ is strongly expansive if there exist $\varepsilon>0$, for any $x,y\in X$, $x\neq y$, we can find a number $n\in\mathbb{N}$, such that $d(T^nx,T^ny)>\varepsilon$.

Def2. $T$ is an expanding map if there exist a constant $c>1$ and a positive number $\varepsilon>0$, for any $x,y\in X$, if $d(x,y)<\varepsilon$, we have $d(Tx,Ty)>cd(x,y)$.

Of course, an expanding map must be weakly expansive. (The definition can be found from Are these two kinds of definitions about expansivity equivalent?)

My quesiton is: must an expanding map be strongly expansive? If not, does there exist a counterexample?

  • $\begingroup$ Consider $x + $ a perturbation that decays at infinity to see why this is false. $\endgroup$ – user296602 Dec 31 '15 at 8:36

They are equivalent if $T$ is injective. Suppose that $\epsilon$ and $c$ witness that $T$ is an expanding map. Let $\delta=\frac{1}2\epsilon$, and let$x$ and $y$ be distinct points of $X$. If $d(Tx,Ty)>\delta$, we're done. Otherwise, $d(Tx,Ty)\le\delta<\epsilon$, and therefore $d(T^2x,T^2y)>cd(Tx,Ty)$. As long as $d(T^nx,T^ny)<\epsilon$, we must have $d(T^{n+1}x,T^{n+1}y)>cd(T^nx,T^ny)$, so by an easy induction we have $d(T^{n+1}x,T^{n+1}y)>c^nd(Tx,Ty)$. Since $c^n$ increases without bound as $n$ increases, at some point we must have $d(T^nx,T^ny)>\delta$, and it follows that $\delta$ witnesses the strong expansiveness of $T$.

Otherwise, however, they are not. Let $X=\{x\in\Bbb R:|x|\ge1\}$ with the usual metric. Let $Tx=2|x|$ for $x\in X$. Let $x,y\in X$ with $x\ne y$. If $0<d(x,y)<1$, then $d(Tx,Ty)>\frac{3}2d(x,y)$, so $T$ is an expanding map. In this answer, however, I showed that $T$ is not strongly expansive.

  • $\begingroup$ Why $Tx\neq Ty$? $\endgroup$ – James Chan Dec 31 '15 at 11:45
  • $\begingroup$ @James: Oversight on my part: I'm used to dealing with injective maps in the context of dynamical systems and unconsciously assumed that $T$ was injective. I'm headed for bed now; when I'm up again I'll either fix it or delete it. $\endgroup$ – Brian M. Scott Dec 31 '15 at 12:15
  • $\begingroup$ @James: As it turned out, I realized that I already had the fix. $\endgroup$ – Brian M. Scott Dec 31 '15 at 12:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.