🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: sci.math.num-analysis
Article #464

"Difficult" oscillatory integrations

#464
From: Massimo Campostrini
Date: Sat, 05 Sep 1992 08:46
26 lines
1486 bytes
   I often have to perform numerical integrals of the form


          /\ infinity
         / 
        /      f(x) dx
       /
     \/ 0

where f(x) is oscillating and the integral of |f(x)| diverges
[f(x) often behaves like sin(omega*x+phi)/sqrt(x) for large x].
Are there any really accurate numerical integration methods for
this class of problems (I need at least 7 digit precision),
apart from rotating the integration contour in the complex plane
to imaginary x?  (contour integration works very well, but takes
a lot of analytical effort: I have identify all poles and cuts
of f, which is a very complicated transcendent function,
depending on parameters...)

   Thanks in advance,

-- 
        Massimo Campostrini, 
Istituto Nazionale di Fisica Nucleare, Sezione di Pisa,
Piazza Torricelli 2, I-56126 Pisa, Italy  ||  Phone: (+39)(50)42093
Internet: campo@sunthpi3.difi.unipi.it    ||  Fax:   (+39)(50)48277
Technical Details
Message-ID: <1992Sep5.094614.450@cine88.cineca.it>
Path: arc1-ca.pugleaf.net!sparky!uunet!spool.mu.edu!sdd.hp.com!usc!rpi!ghost.dsi.unimi.it!37.1!sunthpi3.difi.unipi.it!campo

Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!spool.mu.edu!sdd.hp.com!usc!rpi!ghost.dsi.unimi.it!37.1!sunthpi3.difi.unipi.it!campo
Newsgroups: sci.math.num-analysis
Subject: "Difficult" oscillatory integrations
Message-ID: <1992Sep5.094614.450@cine88.cineca.it>
From: campo@sunthpi3.difi.unipi.it (Massimo Campostrini)
Date: 5 Sep 92 09:46:13 +0100
Reply-To: campo@sunthpi3.difi.unipi.it
Distribution: world
Organization: Istituto Nazionale di Fisica Nucleare, Pisa, Italy
Keywords: numerical integration
Nntp-Posting-Host: sunthpi3.difi.unipi.it
Lines: 26