Another Integration Formula

  • #1
gleem
Science Advisor
Education Advisor
2,602
2,056
According to the author of this article, this integration formula is well-known but rarely taught. Do you know it?

$$\int f(x)dx = xf(x)-\int_{x_{0}}^{f(x)}f^{-1}(t)dt $$

where x0 is a constant and f-1(x) is the inverse function of f(x).
 
  • Like
Likes Demystifier
Physics news on Phys.org
  • #2
Is there a typo somewhere? I get for ##f(x)=e^x##
\begin{align*}
\int e^x\,dx &\stackrel{?}{=} xe^x-\int_{x_{0}}^{e^x}\log(t)\,dt\\
&= xe^x-\left[t\log(t)-1\right]_{x_{0}}^{e^x}\\
&=xe^x-(e^x x-1)+(x_0\log(x_0)-1)\\
&=x_0\log(x_0)
&\neq e^{x}
\end{align*}
Where am I wrong?
 
  • #3
gleem said:
According to the author of this article, this integration formula is well-known but rarely taught. Do you know it?

$$\int f(x)dx = xf(x)-\int_{x_{0}}^{f(x)}f^{-1}(t)dt $$

where x0 is a constant and f-1(x) is the inverse function of f(x).
It's a combination of parts and substitution. You can start from the integral of an inverse function. With ##f(x) = t##, do a change of variables, followed by parts.
$$\int_{t_0}^{t_1}f^{-1}(t)dt = \int_{x_0}^{x_1}xf'(x)dx$$$$ =\bigg[xf(x)\bigg]_{x_0}^{x_1} - \int_{x_0}^{x_1}f(x)dx$$
 
  • Like
Likes PAllen
  • #4
fresh_42 said:
Is there a typo somewhere? I get for ##f(x)=e^x##
\begin{align*}
\int e^x\,dx &\stackrel{?}{=} xe^x-\int_{x_{0}}^{e^x}\log(t)\,dt\\
&= xe^x-\left[t\log(t)-1\right]_{x_{0}}^{e^x}\\
&=xe^x-(e^x x-1)+(x_0\log(x_0)-1)\\
&=x_0\log(x_0)
&\neq e^{x}
\end{align*}
Where am I wrong?
The integral of ##\log t## is ##t\log t - t##.
 
  • Like
Likes Demystifier and fresh_42
  • #5
PeroK said:
It's a combination of parts and substitution. You can start from the integral of an inverse function. With ##f(x) = t##, do a change of variables, followed by parts.
$$\int_{t_0}^{t_1}f^{-1}(t)dt = \int_{x_0}^{x_1}xf'(x)dx$$$$ =\bigg[xf(x)\bigg]_{x_0}^{x_1} - \int_{x_0}^{x_1}f(x)dx$$
You can also just take the derivative of the RHS, and verify that it is f(x). I note that I have never seen this before. It looks cute, but I wouldn't know how useful it is without playing with some examples myself.

[The derivative comes out ##f(x)+xf'(x)-xf'(x)## ]
 
  • Like
Likes PeroK
  • #6
PAllen said:
It looks cute, but I wouldn't know how useful it is without playing with some examples myself.
This looks like one of those integration identities that might find use on some ubsurd integral challenges (which I enjoy), but probably not much use anywhere else. I would be interested to see if it does show up in someones research somewhere.
 
  • #7
Mondayman said:
This looks like one of those integration identities that might find use on some ubsurd integral challenges (which I enjoy), but probably not much use anywhere else. I would be interested to see if it does show up in someones research somewhere.
It could be useful, but it doesn't have a lot of independent value. It's just the full substitution, followed by parts. In that sense, it's nothing new.
 

Similar threads

Replies
4
Views
2K
Replies
31
Views
3K
Replies
2
Views
2K
Replies
6
Views
2K
Replies
1
Views
2K
Replies
3
Views
1K
Replies
3
Views
783
Replies
16
Views
3K
Replies
20
Views
3K
Back
Top