Re: Integrate vs. NIntegrate

*To*: mathgroup at smc.vnet.net*Subject*: [mg51432] Re: Integrate vs. NIntegrate*From*: Bill Rowe <readnewsciv at earthlink.net>*Date*: Sun, 17 Oct 2004 03:05:24 -0400 (EDT)*Sender*: owner-wri-mathgroup at wolfram.com

On 10/16/04 at 4:21 AM, carlos at colorado.edu (Carlos Felippa) wrote: >Which one is the correct answer? > >f=2+Cos[2 x]+ I*Sin[2 x]; > >Print[ N[Simplify[ Integrate[Log[f],{x,-Pi,Pi}]/(2*Pi)]] >//InputForm]; 0.34657359027997264 + 1.5707963267948966*I > >Print[ NIntegrate[Log[f],{x,-Pi,Pi}]/(2*Pi) >//InputForm]; 0.6931471805599428 + 0.*I > >I believe the second one, have doubts about the first. When I look at the following plots Plot[Im[Log[f]], {x, -Pi, Pi}]; Plot[Re[Log[f]], {x, -Pi, Pi}]; I see the imaginary portion of f seems to have equal areas above and below the x-axis indicating the imaginary portion of the integral should be near 0. And the plot of the real portion of f is everywhere positive over the range of interest. So, it seems to me clear the first result cannot be correct and I would take the second as correct. Additionall from what others have reported in this forum, it seems whenever there is a disagreement between using N[Integrate[... and NIntegrate[... it is N[Integrate[... that produces the incorrect answer. In fact, this seems to pop up so often here that I would always use NIntegrate when I wanted a numerical answer instead of converting the results of Integrate to a number. -- To reply via email subtract one hundred and four