Integration question - use of substitution

AI Thread Summary
The discussion revolves around the integration of the function ∫ dx/√(1-e^(2x)). The initial approach using substitution t = √(1-e^(2x)) is questioned, particularly regarding its validity. Participants clarify that while the substitution theorem typically requires a specific form, substitutions can be flexible if they simplify the integral. There is also a correction regarding the sign in the logarithmic term of the final result. The conversation emphasizes the importance of understanding the conditions under which the integral can be evaluated, especially concerning the domain of the function.
twoflower
Messages
363
Reaction score
0
Hi,

I have to find this one:

<br /> \int \frac{dx}{\sqrt{1-e^{2x}}}<br />

Is this right approach?

<br /> \int \frac{dx}{\sqrt{1-e^{2x}}} = \int \frac{e^{2x} dx}{e^{2x} \sqrt{1-e^{2x}}}<br />

Substitution:
<br /> t = \sqrt{1-e^{2x}}<br />

<br /> dt = - \frac{e^{2x}}{\sqrt{1-e^{2x}}} dx<br />

<br /> e^{2x} = 1 - t^2\\<br />

<br /> \int \frac{e^{2x} dx}{e^{2x} \sqrt{1-e^{2x}}} = - \int \frac{dt}{1-t^2} = - \int \frac{1-t}{1-t^2}dt - \int \frac{t}{1-t^2}dt <br />

Substitution:
<br /> y = 1 - t^2<br />

<br /> dy = -2t dt<br />


<br /> z = 1 + t<br />

<br /> dz = dt<br />

<br /> ... = - \int \frac{dz}{z} - \frac{1}{2} \int \frac{dy}{y} = - \ln \left(1 + \sqrt{1-e^{2x}} \right) - \frac{1}{2} \ln \left(e^{2x} \right) + C<br />

I'm afraid that the first substitution is not ok, but could someone please give me more detailed answer?

Thank you.
 
Physics news on Phys.org
Well,the last sign should be a "+" in the term with "ln of e^{2x}"...

\int \frac{dt}{1-t^{2}} can be don also using the substitution

t=\tanh u

Daniel.
 
Last edited:
Thank you dextercioby,

could you tell me whether the use of the first substitution in my approach is ok?
 
Btw we haven't learned hyperbolic functions.
 
I told u,everything is okay,except for the last sign.

Daniel.
 
It's okay.U could do it by simple fraction decomposition,but u said u needed to do it by substitution.

Daniel.
 
Ok, thank you. But I have one doubt about my approach anyway :wink:

The theorem about substition says we can use the substitution in case we have something like this:

<br /> \int f(g(x)) g&#039;(x) dx<br />

But...I'm afraid this is not exactly my case. My integral just doesn't have this form...
 
This one:

<br /> \int \frac{e^{2x} dx}{e^{2x} \sqrt{1-e^{2x}}}<br />
 
  • #10
Well,look at it this way

x\rightarrow t(x)\rightarrow \frac{e^{2x}}{e^{2x}\sqrt{1-e^{2x}}}

and u see that f(x) is your initial function & g(x) is t(x)=sqrt(1-e^(2x)) ...

U needn't worry about the form of the functions.Just make the substitutions which would provide simpler forms for the integrals.

Daniel.
 
  • #11
Ok, so we're quite free to use substitutions if it helps us. I worried about whether my substitution meets the requirements of the theorem..
 
  • #12
Since u haven't provided the intervals on which u wish to integrate that function,then u can do pretty much everything...

For example,you function can be integrated only on the domain in which

e^{2x}&lt;1​

,so that would set a condition on the variable u want to use as a substitution for "x"...

Daniel.
 

Similar threads

Back
Top