1
$\begingroup$

Let $C$ be a curve defined by $y = f(x)$, and define the vertical reflection over $C$ to be the map $(x,y) \mapsto (x,y')$, where $y' = 2 f(x) - y$. In other words, the vertical distance from $(x,y)$ to $C$ is the same as the vertical distance from $(x,y')$ to $C$.

Also, you can similarly define the horizontal reflection over a curve of the form $x = g(y)$.

Now, say you have two curves $C_1$ and $C_2$, and let $R_1$ be the horizontal reflection over $C_1$ and $R_2$ the vertical reflection over $C_2$. I am interested in dynamical systems generated by maps of the form $F = R_2 \circ R_1$.

I found the following PhD thesis involving a specific class of maps of this form: Jensen - Homoclinic points in the composition of two reflections. This work talks about how if the curves $C_1$ and $C_2$ intersect more than once, then the dynamics has chaotic behavior.

I am however, interested in the converse. If the curves intersect only once, then what extra conditions are needed to ensure the system is NOT chaotic? What conditions on $C_1$ and $C_2$ ensure that the dynamics is integrable? In particular I want to know when the orbits are bounded, and lie on simple closed curves.

$\endgroup$
2
  • $\begingroup$ So just to make sure I understand what you're asking, let $f_1(x)$ and $f_2(x)$ be functions. (Are they supposed be be polynomials?) Then you are looking at the dynamics of the map $$ F(x,y) = \Bigl( 2f_2\bigl(f_1(x)-y\bigr),2f_1(x)-y\Bigr). $$ $\endgroup$ Commented May 23, 2021 at 1:23
  • $\begingroup$ @JoeSilverman No, I do not want to assume they are polynomials. And What you wrote is (almost) $R_1 \circ R_2$, but I think you forgot an extra $-x$ in the first component. It doesn't matter that much, but I want to look at $R_2 \circ R_1$ (the horizontal reflection comes first), so it would be $$F(x,y) = \left( 2f_1(y)-x, 2f_2(2f_1(y)-x) - y \right)$$ $\endgroup$
    – Nick
    Commented May 24, 2021 at 0:36

0

You must log in to answer this question.