Re: PSLQ

*To*: mathgroup at smc.vnet.net*Subject*: [mg80812] Re: PSLQ*From*: David Bailey <dave at Remove_Thisdbailey.co.uk>*Date*: Sun, 2 Sep 2007 02:59:37 -0400 (EDT)*References*: <fbaqv8$qf8$1@smc.twtelecom.net>

Clashton at gmail.com wrote: > I am trying to use Peter Bertok's PSLQ algorithm at > > http://library.wolfram.com/infocenter/MathSource/4263/ > > to check for linear relations between logarithms, but it is not > working properly. > > If I do > > PSLQ[N[{Log[3/2], -Log[2], -Log[4/3]}, 200], 200] > > , the program correctly outputs > {1, 1, -1}. > > However, if I try > > PSLQ[N[{Log[3/2], -Log[5], -Log[4/3]}, 200], 200] > > , instead of outputting a set of three large integers which would give > a linear relation equalling zero correct to 10^(-200) > (which I could interpret to mean there was no true linear > relationship, because of the size of the integers), > it outputs a string of error messages. > > Can anyone tell me how to tweak this program to eliminate these > errors? > > Failing that, is there another Mathematica implementation of PSLQ > which would do the same thing? > > I have searched the archive and there was a link to a Mathematica > program by David Bailey, but the link no longer works. > > Thanks for any help with this, > > Jimmy Mc Laughlin > > PS Please reply also to my e-mail address > Clashton at gmail.com > > The "David Bailey" that you need is David H Bailey (not me)! http://crd.lbl.gov/~dhbailey/ David Bailey http://www.dbaileyconsultancy.co.uk