MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Consider a circle $C$ with radius of $r$, we place $m$ balls(treated as point) randomly on it, and each ball $i$ has the mass $m_i$. We define a function $\varphi:C\rightarrow C$ which maps $x\in C$ to the centroid of those balls with distance less than $l<2\pi r$ to $x$(work on arcs).

My question is that, for any point $x\in C$ whether the sequence $\{\varphi(x),\varphi\circ\varphi(x),\cdots,\varphi^n(x)\}$ terminated at some point $x_0$?(no cycles).

If the circle is not discrete but continuous with a density distribution function, what would be the answer?

(I'm not a English native speaker, so sorry for my poor English)

share|cite|improve this question
up vote 4 down vote accepted

Call a point clockwise (resp. anticlockwise) if $\varphi(x)$ lies in clockwise (resp. anticlockwise) direction of $x$. Note that if $x$ is clockwise then all points on $[x,\varphi(x)]$ are clockwise as well. Now if there would be a cycle all points of the circle would have to be passed over and hence all points of the circle are clockwise (or all points anticlockwise). Consider now a probability density function $f$. Then $x$ clockwise is equivalent to e.g. $g(x)=\int_{x-l}^{x+l} (y-x)f(y)dy>0$. Note that $$0<\int_C g(x)dx=\int_C \int_{x-l}^{x+l} (y-x)f(y)dydx=\int_{C} f(y) \int_{y-l}^{y+l} (y-x)dxdy=0$$ which is a contradiction. Hence there are no cycles. Note that this proof also works for the discrete case.

share|cite|improve this answer
Very nice answer! Thank you @user35593! I didn't notice the points $\in [x,\phi(x)]$ are clockwise also. – Paul Apr 24 '14 at 6:17

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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