7
$\begingroup$

I am working with the function:

$f(x) = 4x^5 - 3x^4 + 2x^3 - x^2 + 3$.

One of the critical points $f'(x) = 0$ occurs when $x=0$. Checking the second derivative, we see that $f''(0) = -2$. This would lead one to the conclusion that the point $x=0$ is a local maximum.

However, looking at the plot of the function, we see that $x=0$ looks more like a saddle: enter image description here

So, I am just wondering how we could conclude it is a saddle point if the second derivative is (incorrectly?) saying it is a local max.

Thanks.

$\endgroup$
  • $\begingroup$ Try plotting it for $-1\le x\lt 1$ $\endgroup$ – AgentS Nov 22 '19 at 0:39
  • $\begingroup$ this is one of those times when looks can be deceiving :) $\endgroup$ – peek-a-boo Nov 22 '19 at 0:42
11
$\begingroup$

You have to zoom in more to understand the whole picture:

At first glance it might seem like a saddle point:

enter image description here

Let’s zoom in more

enter image description here

This is starting to look like more of max than a saddle so let’s zoom in a little bit more:

enter image description here

This is definitely a maximum according to our last picture.

Remember that even though it is a small interval and the function increases almost immediately after it, there are infinite points around $0$ such that they give a lower value for our function which makes $f(0)$ greater than its neighbours.

Source: Desmos Graphing Calculator

$\endgroup$
3
$\begingroup$

The graph can lead to bad evaluation, indeed in this case fince $f'(0)=0$ and $f''(0)<0$ the point is of course a local maximum for the function.

enter image description here

$\endgroup$
2
$\begingroup$

Graphs can deceive. You have $f(0)=3$ and the derivative is $$ f'(x)=20x^4-12x^3+6x^2-2x $$ Nothing here hints to the function having an inflection point at $0$. To the contrary, $$ f'(x)=2x(10x^3-6x^2+3x-2) $$ is positive for $-\delta<x<0$ and negative for $0<x<\delta$ (some $\delta>0$), because the term in parenthesis is $-2$ for $x=0$.

Therefore $0$ is a point of local maximum, which is confirmed by $f''(x)=80x^3-36x^2+12x-2$ and $f''(0)=-2$.

In order that $c$ is an inflection point for a polynomial function $f$, one needs that $c$ is a root of multiplicity greater than one and odd of the polynomial $f(x)-f(c)$: just write the Taylor expansion at $c$ to realize it.

$\endgroup$

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.