Mathematica 9 is now available
Services & Resources / Wolfram Forums
-----
 /
MathGroup Archive
2001
*January
*February
*March
*April
*May
*June
*July
*August
*September
*October
*November
*December
*Archive Index
*Ask about this page
*Print this page
*Give us feedback
*Sign up for the Wolfram Insider

MathGroup Archive 2001

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

Search the Archive

Re: A tough Integral

  • To: mathgroup at smc.vnet.net
  • Subject: [mg28555] Re: A tough Integral
  • From: Richard Easther <easther at physics.columbia.edu>
  • Date: Fri, 27 Apr 2001 03:56:09 -0400 (EDT)
  • Organization: Columbia University
  • References: <200104250530.BAA19306@smc.vnet.net> <9c7n4d$sda@smc.vnet.net>
  • Sender: owner-wri-mathgroup at wolfram.com

On 25 Apr 2001, Daniel Lichtblau wrote:

>bobbym1953 wrote:
>> 
>> Does anyone know how I can get the following Integral to at least 60 places,
>> using Mathematica?
>> 
>> Int(1/(cos(x)+x^2)) between x=0 and x=infinity. Both Integrate and NIntegrate
>> seemed helpless. 
> 
> 60 places?! I wish you luck.

Actually, it can be done.  Break the integral into two pieces, I_1 where
the integral is integrated from O<x<N, and I_2 where we integrate between
N<=x<Infinity.

The answer is then simply I_1 + I_2

If N is not too large, the first integral can be done to the rquired
precision using NIntegrate.

The second integral can be evaluated by noting that the integrand

 1/(x^2(1 + Cos[x]/x^2)) 

can be written as a sum with the general term

 (-1)^n (Cos[x]/x^2)^n /x^2

For integer n, Mathematica can evaluate the indefinite integral over the
general term (although it does not seem to be able to give a general
formula) so the series can be integrated term by term.  Taking enough
terms will give I_2 to arbitrary accuracy.

The trick is to choose N large enough so that you do not need a huge
number of terms in the series that approximates I_2, but not so large that
NIntegrate cannot do the first sub-integral to the desired accuracy. 

If you worked really hard, and extracted the analytic form of the integral
over the general term in I_2, it is just possible you may be able to find
an analytic expression for I_2. If you were really lucky you might then be
able to analytically continue N back to zero, and thus evalute the
integral algebraically.

Richard,
 who after doing this started to wonder why this integral needs to be
worked out to 60 significant figures.






  • Prev by Date: Re: RealTime3D without RealTime3D
  • Next by Date: Re: Getting stylized text with a palette button
  • Previous by thread: Re: A tough Integral
  • Next by thread: Re: A tough Integral