I am given the following function:

$$f: D \rightarrow \mathbb{R} \hspace{2cm} f(x) = \ln \bigg (1 + \sqrt{|x|} -x \bigg )$$

where $D$ is the maximum domain of the function. I have to find the number of local extrema points of this function. What I did was to first make the function look a little cleaner by getting rid of the absolute value:

$$ f(x) = \left\{ \begin{array}{ll} \ln(1 + \sqrt{x} - x) & \quad x \ge 0 \\ \ln(1 + \sqrt{-x}-x) & \quad x < 0 \end{array} \right. $$

I know that we have local extrema points where the derivative of the function is $0$ while the derivative approaching that point from right and left has opposite signs.

We can also have a local extremum in a point where the derivative does not exist (an example being the function $h(x) = |x|$). So naturally I found the derivative of the function:

$$ f'(x) = \left\{ \begin{array}{ll} \ \dfrac{1-2\sqrt{x}}{2\sqrt{x}(1+\sqrt{x}-x)} & \quad x > 0 \\ \\ \dfrac{1-2\sqrt{-x}}{2\sqrt{-x}(1+\sqrt{-x}-x)} & \quad x < 0 \end{array} \right. $$

But I got stuck here. I don't know how to continue. What bothers me is that I do not know the domain $D$ and I don't know how can I find it. If I try to find the values of $x$ for which


I get $x_0= \dfrac{1}{4}$ and $x_1 = -\dfrac{1}{4}$, but if I look at the graph of the function:

enter image description here

I can see that we don't have a local extremum point at $x_1 = -\dfrac{1}{4}$, even thought we do have a local extremum point at $x_0 = \dfrac{1}{4}$. By looking at the graph I also see that we have an extremum point at $x = 0$. I'm guessing that is because of the denominator of the derivative makes it such that the derivative is not defined at $x=0$, so we have an extremum point, but shouldn't we also take into consideration the values of $x$ for which the other term in the denominator of the derivative, $(1+\sqrt{\pm x} - x)$ is not defined? And how would we handle that? And how would I find that we have a local extremum point at $x=0$ analytically, without looking at the graph of the function? Also, isn't that point of the graph that is close to $y=-35$ also a local extremum point?

TL;DR: How can I find all the local extremum points of the given function without looking at the graph? The correct answer is $2$ (according to my textbook).


For a fraction to be zero, the numerator must be zero (with one caveat, discussed at the end of this paragraph). So you want to solve $$ 1 - 2 \sqrt{x} = 0 \text{,} $$ restricted to $x > 0$ and check each solution to see if it also makes the denominator zero. (If it does also make the denominator zero, take a limit approaching that potential solution to see what is really happening.)

This says $\sqrt{x} = 1/2$, so $x = 1/4$. Putting that in the denominator, $2 \sqrt{1/4}(1+\sqrt{1/4} - 1/4) = 5/4$, so we do not need to take a limit. Notice that for $0<x<1$, the denominator of $f'$ is positive, so we need only check the sign of the numerator to the left and right of $1/4$ in $(0,1)$. Let's check the signs of the numerators of $f'(1/16)$ and $f'(1/2)$ : $1 - 2/4$ is positive and $1 - 2/\sqrt{2}$ is negative, so we found a local maximum.

Perform a similar procedure on the $x < 0$ half and find no local extrema on that half.

But you have missed something. What is $f'(0)$ (Is it even defined?) and does $f'$ change signs while crossing $x = 0$?

| cite | improve this answer | |
  • $\begingroup$ But the problem is that if I perform a similar procedure on the $x < 0$ half like you said, I do get a local extremum point at $-\dfrac{1} {4}$ and I don't know what's wrong. I shouldn't get a minimum or maximum at that value. $\endgroup$ – user1502 Dec 16 '19 at 21:14
  • $\begingroup$ @user1502 : $x = -1/4$ is not a solution of $-2\sqrt{-x} - 1 = 0$. $\endgroup$ – Eric Towers Dec 16 '19 at 21:27
  • $\begingroup$ Just one more question: Isn't there also something to consider for a value like $x = \dfrac{3+\sqrt{5}}{2}$ since for that value of $x$ the derivative of $f(x)$ is again not defined (denomiantor = 0, because of the term $1+\sqrt{x}-x$). Wouldn't the function have an extremum at this point also? Just like in the case where $x=0$, the case you told me to look at in the last sentence of your answer. $\endgroup$ – user1502 Dec 17 '19 at 23:54
  • $\begingroup$ @user1502 : No extremum. No change of sign in the derivative as $x$ crosses that value. In fact, the function isn't even defined to the right of that point. $\endgroup$ – Eric Towers Dec 18 '19 at 0:02

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.