Theorem: If $f$ is continuous on the interval $[a,b]$ then $\forall y \in [f(a),f(b)], \space \exists c \in [a,b]$ such that $f(c)=y$

I've seen many proofs using the epsilon-delta definition of continuous maps. So I wanted to prove it using a slightly different approach using the definition of continuous maps as maps that preserve sequential convergence.

Proof attempt:

Define $$A:=\{x \in [a,b] : f(x)<y\}$$

Claim: if $c=\sup A$, then $f(c)=y$

We show it by eliminating other possibilities. Suppose $f(c)<y$. Then let $(x_{n})$ be a sequence with $(x_{n}) \to c$ and $x_{n} >c$ for all $n \in \mathbb{N}$. The continuity of $f$ preserves the sequenial convergence of $(x_{n})$ so $f(x_{n}) \to f(c)$. If $(f(x_{n_{k}}))$ is a monotonic subsequence of $(f(x_{n}))$, then we know that there exists $N \in \mathbb{N}$ such that $\forall n^*\geq N$ where $n^* \in \{n_{k}\}$, we have $f(c)<f(x_{n^*})<y$. But then $x_{n^*} \in A$ but $c= \sup A<x_{n}$ for all $n \in \mathbb{N}$. So $f(c)\geq y$.

On the other hand, if $f(c)>y$, we can similarly construct a sequence with $(x_{n}) \to c$ but $x_{n} \in A$ for all $n \in \mathbb{N}$. We again find $N \in \mathbb{N}$ such that $\forall n^*\geq N$ we have $y<f(x_{n^*})<f(c)$. But then $x_{n^*}$ is an upper bound which impossible since $x_{n}<c$ for all $n \in \mathbb{N}$.

We can conclude $f(c)=y$

  • If $x_n > c$, why $f(x_{n^*}) < y$? If $x_n >c$, then $x_n \notin A$, so $f(x_n)\geqslant y$. Taking the limit, $f(c) \geqslant y$. – xbh Dec 6 at 14:06
  • Alright. My bad. Your reasoning is acceptable. However I would only get $f(x_n) < y$ after some $N$. Why $f(c)<f(x_n)$ as well? – xbh Dec 6 at 14:41
  • So the sequence $(f(x_{n})) \to f(c)$. We can construct a monotonic subsequence of $(f(x_{n}))$ then either $f(c)<f(x_{n_{k}})$ or $f(x_{n_{k}})<f(c)$. If $f(x_{n_{k}})<f(c)$ then $f(x_{n_{k}})<f(c)<y$ but then $x_{n_{k}} \in A$ so $f(c)<f(x_{n_{k}})$. But the assumption is that $f(c)<y$ so we can find $f(x_{n_{k}^*})$ such that $f(c)<f(x_{n_{k}^*})<y$. – Sei Sakata Dec 6 at 14:44
  • I hope its clearer now. – Sei Sakata Dec 6 at 14:47
  • Okay. That is not that obvious, so it might be better if you include these in your final proof. Also I think my reasoning would be quicker. Additionally, for the 2nd case I think your $(x_n)$ should be picked from $A$. $x_n <c$ is clearly not sufficient. – xbh Dec 6 at 14:52

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.