MathGroup Archive 2003

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

Search the Archive

Re: Odd behavior (bug?) with Exp[-x^2} and Erf

  • To: mathgroup at smc.vnet.net
  • Subject: [mg42963] Re: Odd behavior (bug?) with Exp[-x^2} and Erf
  • From: AES/newspost <siegman at stanford.edu>
  • Date: Wed, 6 Aug 2003 03:16:39 -0400 (EDT)
  • References: <bgnib6$1pu$1@smc.vnet.net>
  • Sender: owner-wri-mathgroup at wolfram.com

In article <bgnib6$1pu$1 at smc.vnet.net>,
 Bill Rowe <listuser at earthlink.net> wrote:

> On 8/4/03 at 12:45 AM, siegman at stanford.edu (AES/newspost) wrote:
> 
> > Very reluctant to claim I've found a bug in Mathematica, but the 
> > following cells certainly aren't doing what I expect them to:
> 
> >     f[x_] = (Exp[-5(x + 1)^2] + Exp[-5x^2] + Exp[-5(x - 1)^2])^2
> >     Plot[f[x], {x, 0, 2}, PlotRange -> {{0, 2}, {0, 2}}];
> >     g[y_] = Integrate[f[x], {x, 0, y}, Assumptions -> y > 0]
> >     Plot[g[y], {y, 0.001, 2}, PlotRange -> {{0, 2}, {0, 1}}];
> 
> What do you see that you don't expect?


f[x] as given in first line and plotted in second line is everywhere 
positive.

g[y] in third line gives a complicated expression with a lot of Erf's in 
it (appended below), and no error messages or warnings.

g[y] as plotted in 4th line doesn't increase monotonically --rises, then 
falls, then rises again --  even though it's supposedly the integral 
over an increasing range of an everywhere positive function.

I suspect some of the Erf's have been evaluated without proper care as 
to whether the limits are positive or negative, but I haven't dug in to 
do the grunt work of doing all the integrations carefully by hand and 
checking on the Mathematica result.


g[y]\!\(\(\(1\/\(2\ \[ExponentialE]\^10\)\)\((\@\(\[Pi]\/10\)\ 
\((\(-2\)\ \
\[ExponentialE]\^\(15/2\)\ Erf[\@\(5\/2\)] - \[ExponentialE]\^10\ 
Erf[\@10] + 
          2\ \[ExponentialE]\^\(15/2\)\ Erf[\@\(5\/2\)\ \@\(1\/\((1 - 2\ 
y)\)\
\^2\)\ \((1 - 2\ y)\)\^2] + \[ExponentialE]\^10\ Erf[\@10\ 
\@\(1\/\((\(-1\) + \
y)\)\^2\)\ \((\(-1\) + y)\)\^2] + 
          2\ Erf[\@10\ y] + \[ExponentialE]\^10\ Erf[\@10\ y] + \
\[ExponentialE]\^10\ Erf[\@10\ \@\(1\/\((1 + y)\)\^2\)\ \((1 + y)\)\^2] 
+ 
          2\ \[ExponentialE]\^\(15/2\)\ Erf[\@\(5\/2\)\ \@\(1\/\((1 + 2\ 
y)\)\
\^2\)\ \((1 + 2\ y)\)\^2])\))\)\)\)


  • Prev by Date: Re: Re: Background Colors in The Format Menu
  • Next by Date: Re: Page preview in Mathematica ?, Kerning
  • Previous by thread: Re: Odd behavior (bug?) with Exp[-x^2} and Erf
  • Next by thread: Re: Odd behavior (bug?) with Exp[-x^2} and Erf