Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.soft-sys.math.mathematica > #9958

Re: Integration Problem

From Oliver Jennrich <oliver.jennrich@gmx.net>
Newsgroups comp.soft-sys.math.mathematica
Subject Re: Integration Problem
Date 2012-05-04 10:26 +0000
Organization Steven M. Christensen and Associates, Inc and MathTensor, Inc.
Message-ID <jo0as0$c8u$1@smc.vnet.net> (permalink)
References <jntg1g$g$1@smc.vnet.net>

Show all headers | View raw


Michael Musheghian <michael.musheghian@gmail.com> writes:

> Greetings!
>
> I found that evaluation of this 2 integrals yield a bit different result. What could be the reason?
>
> Integrate[E^(-1/10 ((1 + r2z)^2)), {r2z, -Infinity, Infinity}]
>
> Integrate[E^(-0.1 ((1 + r2z)^2)), {r2z, -Infinity, Infinity}]

Numerics. The first integral evaluates symbolically, the second one
semi-numerically.
Mathematica 8 yields Sqrt[10 \[Pi]] for the former and 
5.60499 - 2.03152*10^-16 I for the latter. 

You can avoid the very small imaginary part by either calculating the
equivalent integral 

Integrate[E^(-0.1 ((r2z)^2)), {r2z, -Infinity, Infinity}] 

(i.e. performing a shift in the integration variable) or by having Mathematica calculate the integral fully symbolically:

Integrate[E^(-a ((r2z)^2)), {r2z, -Infinity, Infinity}, 
 Assumptions -> {a > 0}]

which yields not surprisingly Sqrt[\[Pi]]/Sqrt[a]


-- 
Space - The final frontier

Back to comp.soft-sys.math.mathematica | Previous | Next — Previous in thread | Find similar | Unroll thread


Thread

Integration Problem Michael Musheghian <michael.musheghian@gmail.com> - 2012-05-03 08:36 +0000
  Re: Integration Problem Oliver Jennrich <oliver.jennrich@gmx.net> - 2012-05-04 10:26 +0000

csiph-web