Let $\Omega_1 \supset \Omega_n \supset\cdots$ a decreasing sequence of bounded, convex and open sets in $R^n$. Define $\Omega = \operatorname{int} \left(\overline{\bigcap \Omega_n}\right)$ and supose that $\Omega \neq \emptyset$ , convex , open and bounded. Then is true that $\partial \left(\bigcap \overline{\Omega_n}\right) = \partial \Omega$?

Intuitively this is true, but i dont know if is true. If is true it will help me understand the proof of a theorem. To prove i am trying to use this : $\operatorname{int}(\overline{A}) = \overline{A}$ for every bounded convex set in $R^n.$

someone can give me a help to prove or disprove what i am asking?

thanks in advance

  • $\begingroup$ where you have the closure in each case it changes, should it be the same in each case? $\endgroup$ – Ellya May 7 '14 at 19:22
  • $\begingroup$ the closures are in the manner that i writed =) $\endgroup$ – math student May 7 '14 at 19:36

Yes, this is true. Recall that a convex set has nonempty interior if and only if it is not contained in any hyperplane. Since the closure of $\bigcap \Omega_n$ has nonempty interior, it is not contained in a hyperplane. Hence, $\bigcap \Omega_n$ is not contained in a hyperplane. So it has an interior point. We may assume it's the origin; so, $0\in \Omega_n$ for all $n$.

By convexity, every $\Omega_n$ can be described as $\Omega_n = \{ t\xi: 0\le t<\rho_n(\xi) \} $ where $\xi$ runs over all unit vectors. Let $\rho(\xi) = \inf_n \rho_n(\xi)$. It is routine to verify that: $$ \{ t\xi: 0\le t<\rho (\xi) \} \subseteq \bigcap_n \Omega_n \subseteq \{ t\xi: 0\le t\le \rho (\xi) \} $$ hence, $$ \overline{\bigcap \Omega_n} = \{ t\xi: 0\le t\le \rho (\xi) \}$$ and $$\Omega = \{ t\xi: 0\le t<\rho (\xi) \},\quad \partial \Omega = \{ t\xi: t = \rho (\xi) \}$$ Along the same lines, $$ \bigcap \overline{\Omega_n} = \{ t\xi: 0\le t\le \rho (\xi) \} $$ which has the boundary $$ \partial \left(\bigcap \overline{\Omega_n}\right) = \{ t\xi: t = \rho (\xi) \} = \partial\Omega $$


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.