Challenging Integration
- To: mathgroup at smc.vnet.net
- Subject: [mg12747] Challenging Integration
- From: quartier at umnw.ethz.ch (Robin Quartier)
- Date: Wed, 10 Jun 1998 03:04:10 -0400
- Organization: ETH Zurich
- Sender: owner-wri-mathgroup at wolfram.com
Good Morning everybody.
I have the following distribution:
epsilon[x_]:=(a*b/(b^2 - a^2))*(b*Exp[-a*x] - a*Exp[-b*x])
a,b >0
This function is kind of symmetric, in the sens that it is stable for
the permutation of a and b.
It is a distribution, normalised on the positive half-axis
The integration I want to make is:
Integrate[epsilon[x]*Log[epsilon[x]],{x,0,Infinity}]
Mathematica does not want to do it.
I can get a primitive,
Integrate[epsilon[x]*Log[epsilon[x]],x]
But I don't think that the output I get is correct, because it is not
symmetricalfor the permutation of a and b anymore. Namely, it contains
an Hypergeometric function that diverges for a>b, and converges for
a<b, which can not be.
Ideas, questions and comments welcomed!
Best regards
Robin Quartier