Re: NIntegrate
- To: mathgroup at smc.vnet.net
- Subject: [mg36527] Re: NIntegrate
- From: "Rob_jack" <rob_jackNSP at libero.it>
- Date: Wed, 11 Sep 2002 03:27:47 -0400 (EDT)
- References: <alkhi3$ve$1@smc.vnet.net>
- Sender: owner-wri-mathgroup at wolfram.com
"Rob_jack" <rob_jackNSP at libero.it> wrote
There is an error in the previous message.
This is the just msg:
**********************************************
In[1]=f[y_,z_]: = (y^4/( (1+(8.44*10^-4)^2 * (1+z)^2 y^2) (Exp[y]+1) )
In[2]=F[z_]: = NIntegrate[f[y,z], {y, 0, Infinity}
**********************************************
Mathematica 4.0 says:
****************************************************************************
NIntegrate: : inum : Integrand 1.07577/( 1+7.12336*10^-7 (1. + z)^2 ) is not
numerical at {y}={1.}.
****************************************************************************
What is it??
Moreover, I have used the following procedure:
******************************************************
If z<<1,
f[y,z]=Sum[(-1)^n*a^n*(1+z)^2n *(y^(2n+4)/(Exp[y]+1),{n, 0, N} ]
here a=(8.44*10^-4)^2
therefore:
F[z]:=Sum[(-1)^n*a^n*(1+z)^2n *(y^(2n+4)/(Exp[y]+1) * Gamma[2 n+5]*Zeta[2
n+5],{n, 0, N} ]
this series (N<+oo) only approximates the function for small z, and me it
interests the behavior of F for z in the range 800-1300.
Thx in advance.
Rob_jack