I had an idea for a proof of the Extreme Value Theorem, and I was wondering if it was valid. Any advice you have would be greatly appreciated. Thank you!

Prove: If a function f(x) is continuous on a closed interval [a,b], then f(x) has both a maximum and a minimum on [a,b]

Proof: Case 1: f is constant on [a,b]

If f is constant, then for any $x_1$ < $x_2$, f($x_1$) = f($x_2$). Therefore, f($x_1$) ≥ f($x_2$) and f($x_1$) ≤ f($x_2$). So, by definition of maxima and minima, f has at least one of both on [a,b]. This satisfies the theorem.

Case 2: f is strictly increasing or strictly decreasing on [a,b]:

Let's start with strictly increasing. By definition, for each $x_1$ < $x_2$, f($x_1$) ≤ f($x_2$). Because x < b for any x in (a,b), f(x) ≤ f(b). so the endpoint f(b) is a maximum on f on [a,b]. We also know that, because a < x for any x in (a,b), f(a) ≤ f(x). So f(a) is a minimum on [a,b]. This satisfies the theorem. The case for a strictly decreasing function f on [a,b] is similar.

Case 3: f is both strictly increasing and strictly decreasing on [a,b]

Divide the open interval (a,b) into smaller subintervals $U_i$ and $D_i$. On $U_i$, f is strictly increasing and possibly constant, and on $D_i$, f is strictly decreasing and possibly constant. Construct the subintervals so that, if there is a constant section of the graph in a subinterval, it is near the right endpoint of the respective subinterval. Each $U_i$ is followed by a $D_i$, and vice versa.

Assume the function has at least one strictly increasing interval. By the argument for case 2, at the right endpoint of all subintervals $U_i$, the value will be a maximum $M_i$. If f is constant near the right endpoint, the right endpoint is still a max for the total interval $U_i$ by the definition of a maximum. Because each interval $U_i$ is followed by an interval $D_i$, and because f is strictly decreasing on the neighboring $D_i$, $M_i$ will be a maximum for the combined interval $D_i$ + $M_i$. Adding all combined intervals $D_i$ + $M_i$ gives (a,b), so the largest $M_i$ found will be the total maximum/maxima of f on (a,b). Let S = {$M_i$ | $i∈ℕ$, $M∈f$ }. Because every continuous function is bounded, the maxima of f must also be bounded, so S is bounded above. We know S is non-empty because we assume there is at least one strictly increasing interval $U_i$, and therefore at least one $M_i$ in (a,b). Therefore by the Supremum Axiom, there must be a $M_{max}$$M_i$ for all maxima in S. The endpoints f(a) and f(b) may be greater than the $M_{max}$, so the endpoints must be compared with $M_{max}$, but no matter the outcome of the comparison, there will still be at least one maximum for f on [a,b].

So, f has at least one maximum on [a,b] if it is strictly increasing on at least one subinterval of (a,b). The proof for f having at least one minimum on [a,b], assuming it is strictly decreasing on at least one subinterval of (a,b), is similar.

This concludes the proof.


I googled this:

Nowhere Monotonic Continuous Function


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.