How to test if a definite integral is finite or not?

Click For Summary
To determine if a definite integral is finite, particularly for integrals over infinite domains, the behavior of the integrand as x approaches infinity is crucial. A necessary condition is that the integrand must approach zero; however, this alone is not sufficient for finiteness. The discussion highlights that a positive or negative function will have a finite integral if it decreases faster than 1/x as x approaches infinity. Additionally, the analysis must consider points where the function diverges, not just its behavior at infinity. Understanding these conditions is essential for evaluating integrals with complex functions and parameters.
arroy_0205
Messages
127
Reaction score
0
Suppose I have a complicated integral whose exact evaluation seems extremely difficult or may be even impossible, in such a case is there any way to tell if the integration result is finite or not? suppose the problem is
<br /> \int_{-\infty}^{\infty} f(x;a,b) dx<br />
I think there might be some conditions on the function involved. Actually the function contains some parameters also (a,b) which can be taken to be constants for a particular case. Now I am looking for a condition general enough to handle arbitrary parameters, ie, can I tell if the integral is finite for any arbitrary values of those parameters? If yes, then how or under what condition? Take as an example:
<br /> \int_{-\infty}^{\infty} e^{-2a \tanh^2(bx)} dx<br />
Is this finite?
 
Physics news on Phys.org
I may be a little rusty. If I remember correctly, |tanh(bx)| -> 1 as x -> oo. In which case, the integral certainly is NOT finite. A necessary condition for integral to be finite with an infinite domain is the integrand go to zero as x -> oo. Also, this condition is not sufficient!
 
If i remember correctly, besides using definition you can use the fact that any positive (or negative) function has a finite integral, if its falling with a greater exponent than 1/x. (equivalently - if its below x^-(1+eps) for any x greater than some x0). It doesn't work ok for an alternating function.
This applies if you are interested in a true integral, not prime value. Also, this covers just what happens when x goes to infinity, not points where function itself diverges.
 

Similar threads

  • · Replies 2 ·
Replies
2
Views
1K
  • · Replies 16 ·
Replies
16
Views
4K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 6 ·
Replies
6
Views
3K
  • · Replies 4 ·
Replies
4
Views
1K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 11 ·
Replies
11
Views
2K