MathGroup Archive 1998

[Date Index] [Thread Index] [Author Index]

Search the Archive

probs with InverseFourierTransform and Integrate from v2.2 to v3


  • To: mathgroup@smc.vnet.net
  • Subject: [mg10848] probs with InverseFourierTransform and Integrate from v2.2 to v3
  • From: Alia Atlas <akatlas@cs.bu.edu>
  • Date: Tue, 10 Feb 1998 21:01:59 -0500
  • Organization: Boston University

I'm having a problem doing convolutions on mathematica 3.0 on an SGI,
although the same code works on mathematica v2.2 on a sparc

The code is (from a clean start of mathematica):

In[1]:= <<Calculus`FourierTransform` In[2]:=
<<Statistics`ContinuousDistributions` In[3]:= dist1 =
NormalDistribution[10,3] Out[3]= NormalDistribution[10, 3]
In[4]:= pdf1 = PDF[dist1, x]

                     1
Out[4]= ---------------------------
                    2
           (-10 + x) /18
        3 E              Sqrt[2 Pi]

In[5]:= trans1 = FourierTransform[pdf1, x, s]
                      2
         10 I s - (9 s )/2
Out[5]= E

In[6]:= rev1 = InverseFourierTransform[trans1, s, x]
                                              2
                                 10 I s - (9 s )/2 Out[6]=
InverseFourierTransform[E                 , s, x] In[7]:=
NIntegrate[rev1, {x, 0, 30}]

NIntegrate::inum: 
                                                           2
                                            10. I s - 4.5 s
   Integrand InverseFourierTransform[2.71828                , s, 15.]
     is not numerical at {x} = {15}.

Out[8]= NIntegrate[rev1, {x, 0, 30}]


Whereas, in v2.2, I get an actual numerical answer.

Can anyone help here?

Thanks,
Alia Atlas



  • Prev by Date: Re: nasty integrals that Mathematica can't do
  • Next by Date: Error Bars
  • Prev by thread: Mathematica 3.0.2, NeXT Intel
  • Next by thread: Error Bars