2
$\begingroup$

   Let $\lbrace a_{n}\rbrace_{n=1}^{\infty}$ be a strictly increasing bounded sequence of real numbers such that $\lim\limits_{n \to \infty}$ $a_{n}=A$. Let $f:[a_{1},A]\rightarrow \mathbb{R}$ be a continuous function such that for each positive integer $i$, $f\vert_{[a_{i},a_{i+1}]}:[a_{i},a_{i+1}] \rightarrow \mathbb{R}$ is either strictly increasing or strictly decreasing. \       Consider the set\       $$B = \left\lbrace M \in \mathbb{R} \middle|\text{ there exist infinitely many }x \in [a_{1},A]\text{ such that } f(x)=M\right\rbrace.$$ Then prove that the cardinality of $B$ is atmost one.

If the function is strictly increasing throughout its domain then it is injective and hence $B$ is empty. Similarly if it’s decreasing throughout. To have infinite pre images for a single point, the function must have infinite bumps and geometrically it looks that at most one such point is possible. How do we prove it rigorously? Please help.

$\endgroup$
1
  • $\begingroup$ Have you tried proving that, if $M\in B$, then $M=f(A)$? $\endgroup$ – bof Feb 16 '20 at 0:56
3
$\begingroup$

Hint: if $M\in B$ then there is a sequence $(y_n)\subseteq [a_1,A]$ such that $y_i\neq y_j$ for all $i\neq j$ and such that $f(x_i)=M$ for $i\in \mathbb N$. Since $f$ is injective on each $[a_i,a_{i+1})$, each of these subintervals contains at most one $y_i$, so there are subsequences $(a_{n_k})$ and $(y_{n_k})$ such that $a_{n_k}\le y_{n_k}\le a_{n_{k+1}}$, from which it follows that $y_{n_k}\to A$. Continuity of $f$ implies that $f(A)=M$.

$\endgroup$
4
  • $\begingroup$ You have talked about a sequence yn and about xi which I am unable to comprehend. Can you please explain that. Thank you. $\endgroup$ – Lawrence Mano Feb 16 '20 at 19:47
  • $\begingroup$ If $M\in B$ then there are infinitely many $x\in [a_1,A]$ s.t.$f(x)=M$ But since $f$ is injective on $I_i=[a_i,a_{i+1})$ no two $x's$ can be in any one $I_i$. Now, since $M\in B$ there are infinitely many $x's$ so there must be at least countably many, say $x_k$. And since no two $x_k$ can be in any one $I_i$, we obtain the subsequences of my answer. $\endgroup$ – Matematleta Feb 18 '20 at 2:07
  • $\begingroup$ Thank you very much for your exposition $\endgroup$ – Lawrence Mano Feb 18 '20 at 14:04
  • $\begingroup$ You're welcome. Glad to help! $\endgroup$ – Matematleta Feb 18 '20 at 22:23

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.