I'm trying to understand a proof for the extreme value theorem: Let $f$ be a continuous real-valued function on a closed interval [a, b]. Then $f$ is a bounded function. Moreover, $f$ assumes its maximum and minimum values on [a, b]; that is, there exist $x_0$, $y_0$ in [a, b] such that $f(x0) ≤ f(x) ≤ f(y0)$ for all $x ∈ [a, b]$.

I understand how to prove that $f$ is bounded, but now I'm trying to understand a proof of showing $f$ assumes its maximum and minimum values. To prove it, let $M = \text{sup}\{f(x):x\in[a,b]\}$ which we know exists because $f$ is bounded. Now, for all $n \in \mathbb{N}$, there exists a sequence $(y_n)$ such that $M - 1/n < f(y_n) \leq M$. Then the limit of $f(y_n) = M$, (but how can I prove this is the limit?) Then, by the Bolzano-Weiersrtass theorem, we can find a subsequence $(f(y_{nk}))$ with limit $y_0 \in [a,b]$. Then since $f$ is continuous, we know that lim$f(y_{nk})=f(y_0)$. Since $(f(y_n))$ is convergent, it has the same limit as its subsequence $(f(y_{nk}))$, $M$. Therefore, lim$f(y_n) = f(y_0) = M$, which is the maximum value and shows that the function assumes its maximum. My only question is how do we know that $(f(y_n))$ converges?

  • $\begingroup$ Because by choice you have $|M-f(y_n)| \le {1 \over n}$. Pick $\epsilon>0$ and $N$ such that ${1 \over N} < \epsilon$. Then for $n \ge N$ we have $|M-f(y_n)| < \epsilon$. $\endgroup$ – copper.hat Mar 24 '17 at 23:42

You know it converges because of the squeeze theorem.

For each $n \in \mathbb{N}$, $M-1/n < f(y_n) ≤ M$. Hence by the squeeze theorem, \begin{align*} M = \lim_{n\rightarrow\infty}M-\frac{1}{n} \leq \lim_{n\rightarrow\infty}f(y_n) ≤ \lim_{n\rightarrow\infty}M = M \end{align*}

Therefore $\lim_{n\rightarrow\infty}f(y_n)$ must exist and is equal to $M$

  • $\begingroup$ That makes sense, why do I need to use a subsequence then? $\endgroup$ – TheStrangeQuark Mar 25 '17 at 2:47
  • $\begingroup$ I have no idea, I wondered about that part too. It looks redundant to me. $\endgroup$ – Harambe Mar 25 '17 at 2:56
  • $\begingroup$ I was thinking it was to show that $y_0$ is in the interval [a,b]. But it still feels redundant. $\endgroup$ – TheStrangeQuark Mar 25 '17 at 3:02
  • 1
    $\begingroup$ Ah it makes sense now - we've shown that $f(y_n)$ converges to $M$, but it's not obvious that $y_n$ is convergent. You want to prove that "for some $y_0 \in [a, b], f(y_0) = M$". Without the subsequence part we have no idea how $y_n$ is behaving - we've just shown that $f$ attains some maximum without talking about where it attains its maximum. $\endgroup$ – Harambe Mar 25 '17 at 3:16

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.