How Is the Second Derivative of an Inverse Function Calculated?

Click For Summary

Discussion Overview

The discussion centers around the calculation of the second derivative of an inverse function, exploring the mathematical derivation and the validity of certain formulas related to this topic. Participants engage with both theoretical aspects and practical implications of the derivatives involved.

Discussion Character

  • Technical explanation
  • Mathematical reasoning
  • Debate/contested

Main Points Raised

  • One participant presents a formula for the second derivative of an inverse function, suggesting that it can be derived from the first derivative using the relationship between the derivatives of x and y.
  • Another participant proposes using substitution to solve for the second derivative, indicating an alternative approach to the problem.
  • A later reply questions the validity of the initial formula and suggests a correction regarding the integral of the inverse function, asserting that the formula presented in a linked Wikipedia page contains an error.
  • There is a discussion about terminology, with one participant correcting another's use of "derivate" to "differentiate," highlighting the importance of precise language in mathematical discourse.
  • Participants express uncertainty about the reasonableness of the proposed deductions and formulas, indicating a need for further clarification and discussion.

Areas of Agreement / Disagreement

Participants do not reach a consensus on the correctness of the formulas presented. There are competing views regarding the derivation and validity of the second derivative of an inverse function, as well as the interpretation of related integrals.

Contextual Notes

Some participants note potential errors in the formulas and the need for clearer explanations of the derivations. There is also mention of the dependency on definitions and the implications of using specific mathematical terms.

Jhenrique
Messages
676
Reaction score
4
I can derivate x(y) wrt y using the derivative of y(x) wrt x, follows the formula: [tex]\frac{dx}{dy}=\frac{1}{\frac{dy}{dx}}[/tex] until same the 2nd derivative (taking the 2nd diff form of x and deriving wrt to x):[tex]d^2x=\frac{d^2 x}{dy^2} dy^2 + \frac{dx}{dy} d^2y[/tex] [tex]\frac{d^2x}{dx^2}=\frac{d^2 x}{dy^2} \frac{dy^2}{dx^2} + \frac{dx}{dy} \frac{d^2y}{dx^2}[/tex] [tex]0=\frac{d^2 x}{dy^2} \frac{dy^2}{dx^2} + \frac{dx}{dy} \frac{d^2y}{dx^2}[/tex] solving for d²x/dy²: [tex]\frac{d^2x}{dy^2}=-\frac{d^2y}{dx^2}\frac{1}{\left( \frac{dy}{dx} \right)^3}[/tex] I think that this is a razoable deduction for the formula of d²x/dy² in terms of derivatives of y wrt x.


Now, where it came from this formula?

980a4d9f71d1ca064a5193fedcde3d0d.png



links:
http://en.wikipedia.org/wiki/Integration_of_inverse_functions
http://en.wikipedia.org/wiki/Inverse_functions_and_differentiation
 
Physics news on Phys.org
Put y=f(x) and solve by substitution method.
 
Jhenrique said:
I can derivate x(y) wrt y using the derivative of y(x) wrt x, follows the formula: [tex]\frac{dx}{dy}=\frac{1}{\frac{dy}{dx}}[/tex] until same the 2nd derivative (taking the 2nd diff form of x and deriving wrt to x):[tex]d^2x=\frac{d^2 x}{dy^2} dy^2 + \frac{dx}{dy} d^2y[/tex] [tex]\frac{d^2x}{dx^2}=\frac{d^2 x}{dy^2} \frac{dy^2}{dx^2} + \frac{dx}{dy} \frac{d^2y}{dx^2}[/tex] [tex]0=\frac{d^2 x}{dy^2} \frac{dy^2}{dx^2} + \frac{dx}{dy} \frac{d^2y}{dx^2}[/tex] solving for d²x/dy²: [tex]\frac{d^2x}{dy^2}=-\frac{d^2y}{dx^2}\frac{1}{\left( \frac{dy}{dx} \right)^3}[/tex] I think that this is a razoable deduction for the formula of d²x/dy² in terms of derivatives of y wrt x.
I don't know if the above is reasonable (no such word as "razoable" in English). It would help if you added some words to say what you're doing.

BTW, you don't "derivate" something to get its derivative - you differentiate it. As far as I know, "derivate" is not a word, certainly not as used in mathematics.
Jhenrique said:
Now, where it came from this formula?

980a4d9f71d1ca064a5193fedcde3d0d.png
There seems to be an error in this formula.
It should say the following:
$$ \int f^{-1}(y)dy = yf^{-1}(y) - F ° f^{-1}(y) + C$$
The error in the wiki page you linked to is in the first term on the right side. They have x f-1(y), but it should by yf-1(y). The three examples are consistent with what I have, but not what they have for their general formula.

Calculating this integral is not very difficult using integration by parts.

Let y = f(x) (⇔ x = f-1(y))
Then dy = f'(x) dx

Using this substitution, the integral ##\int f^{-1}(y)dy ## becomes ##\int x \cdot f'(x)dx##.

Using integration by parts, with u = x and dv = f'(x)dx,
we have du = dx and v = ∫f'(x)dx = f(x). (Don't need the constant yet.)

So ##\int x \cdot f'(x)dx = x f(x) - \int f(x) dx##
Since x = f-1(y), and f(x) = y, we have
f-1(y) * y + F(x) + C

Here, F is an antiderivative of f.
Jhenrique said:
 

Similar threads

  • · Replies 6 ·
Replies
6
Views
4K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 1 ·
Replies
1
Views
3K
Replies
1
Views
1K
  • · Replies 8 ·
Replies
8
Views
3K
  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 10 ·
Replies
10
Views
3K
  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 4 ·
Replies
4
Views
2K