First, Second Third order derivatives and inflection point

161 Views Asked by At

$1.$ If $\,f'(a) \neq 0\, \text{ and } \,f''(a) = 0\,$, then there is a non-stationary point of inflection at $\,x = a\,$.To confirm, you check the sign of $\,f''(x)\,$ for values of $x$ on the left and right of $a$. If the sign changes, that means concavity changes, and thus, $a$ is the non-stationary point of inflection.

$2.$ If $\,f'(a) = 0\, \text{ and } \,f''(a) = 0\,$, then there is a stationary point of inflection at $\,x = a\,$. To confirm, you check the sign of $\,f''(x)\,$ for values of $x$ on the left and right of $a$. If the sign changes, that means concavity changes, and thus, $a$ is the stationary point of inflection.

Confusion 1: Are $1$ and $2$ correct statements?

Confusion 2: It's said that if $\,f''(a) = 0\, \text{ and } \,f'''(a) \neq 0\, (i.e., \,f'''(a)\,$ is actually some well-defined number and that number is not $0$), then there is an inflection point at $x=a$. My dilemma is that shouldn't it be as $\,f'''(a) =0 \,$ and if there is a change of sign in $f'''(x)$ for values of $x$ on the left and right of $a$, then $x=a$ can be termed as the point of inflection? Also, will this inflection point always be non-stationary?

Additionally, I would appreciate it if you could demonstrate the $f'(x)$, $f''(x)$, and $f'''(x)$ at $x=a$ through graphs and then define the point of inflection.

I have looked at the comments and since the length of my response doesn't fit the comments, I have put it below:

As $f ' ' '(a)≠0$ so $f ' ' '(a)>0$ or $f ' ' '(a)<0$ and that in turn means $f ' '(a)>0$ or $f ' '(a) <0$ around the point $x=a$. This means $f' '$ is strictly monotonic but it is either positive on both sides of $x=a$ or negative on both sides of $x=a$. There will be no change in the sign of $f ' '$ but concavity will change at $x=a$. That is, if you consider $f $ as the function then slope of $f $ will be $f '$ and at point of inflection $f ' '$ will be equal to 0. And in the immediate neighbourhood of the inflection point(both sides), slope ( in this case $f ' $) will either go from less increasing to more increasing OR less decreasing to more decreasing and thus the change in concavity. But it will never change sign. Now consider $f '$ as the function and its slope as $f ' '$ and then $ f ' ' '$ as the requirement for point of inflection. Following what is stated above, if $x=a$ is an inflection point then change in slope (i.e. $ f ' ' '$) will be zero at $x=a$ but it will either be positive on both sides of $a$ or negative on both of $a$ i.e slope $ f ' '$ will either go from less increasing to more increasing OR less decreasing to more decreasing.

So, shouldn't $f ' ' '(a)=0$ if $x=a$ is an inflection point rather than $f'(a) \neq 0$? I am still confused and look forward to hearing from you all. I hope I have been clear in putting up my dilemma.

1

There are 1 best solutions below

0
On

"Stationary" means $f'(a) = 0$. Nothing more or less.

An inflection point is a point where the concavity of the function changes direction. For twice differentiable functions, $f'' > 0$ where the function is concave up, and $f'' < 0$ where the function is concave down. $f''$ can only switch from one to the other by passing through $0$ (even without assuming $f''$ is continuous, because it is a derivative). Thus at an inflection point $a, f''(a) = 0$.

But it is possible for $f''(a) = 0$ without $a$ being an inflection point. $y = f''(x)$ could touch the line $y = 0$, but then turn around and leave in the same direction from which it came, like $y = x^2$ does. But that would mean that $a$ was a local extremum of $f''$, and therefore that the derivative of $f''$ is $0$. I.e., if $f''(a) = 0$, but $a$ is not an inflection point, then $f'''(a) = 0$ as well.

Thus if $f'''(a) \ne 0$, we know this doesn't occur, and thus we know that $a$ must be an inflection point even though we haven't directly checked that $f''$ has opposite signs on either side of $a$.

So, to find where a thrice-differentiable $f$ has inflection points, you find each of the points $a$ where $f''(a) = 0$, as these are the candidate values where inflection points might occur. To determine if they actually are inflection points, you have two choices:

  • Choose points on either side of each candidate, between it and any other candidate points, and evaluate $f''$ at those points. If the results have the same sign on either side of the candidate being tested, then it is not an inflection point. But if they were of different signs, it is an inflection point.
  • Evaluate $f'''$ at each candidate point. If $f'''$ is not zero there, the candidate is an inflection point. If $f'''$ is zero there, then you have to resort to the first method. It might still be an inflection point, or it might not.