Path: g2news2.google.com!news2.google.com!border1.nntp.dca.giganews.com!border4.nntp.dca.giganews.com!border2.nntp.dca.giganews.com!nntp.giganews.com!news-out.octanews.net!mauve.octanews.net!news.astraweb.com!border6.newsrouter.astraweb.com!not-for-mail
Newsgroups: sci.math.num-analysis,comp.theory
From: Tim Little <t...@little-possums.net>
Subject: Re: insuring convergence of an algorithm (solving a set of equations)
References: <65649bb5-3368-4906-847e-fef9d6781941@b8g2000vbi.googlegroups.com>
 <4d67c689.1089798414@text.giganews.com>
 <cfacd7b7-bdb2-4381-9f38-00df5fabd1c0@t8g2000vbd.googlegroups.com>
 <4d680719.1106327001@text.giganews.com>
 <d62255fb-d647-497b-91e4-09f2d6cd3a06@r4g2000vbq.googlegroups.com>
 <4d681764.1110498889@text.giganews.com>
 <4ad78f46-d9dc-4df2-85c3-5e82339d4b87@w36g2000vbi.googlegroups.com>
 <slrnimgids.dj1.tim@soprano.little-possums.net>
 <7d5ce514-7ffb-40f5-9f31-81fcca57967d@by6g2000vbb.googlegroups.com>
User-Agent: slrn/pre0.9.9-111 (Linux)
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <slrnimk58n.dj1.tim@soprano.little-possums.net>
Date: 27 Feb 2011 09:10:47 GMT
Lines: 16
Organization: Unlimited download news at news.astraweb.com
NNTP-Posting-Host: 12f6fb26.news.astraweb.com
X-Trace: DXC=?EP[MoG=?;k:LXj0CA`3KmL?0kYOcDh@jSBc;\8ijUdk4kf8IeZE9Qm7@`nVD75_Xn9o96_A>OC;a

On 2011-02-26, pamela fluente <pamelaflue...@libero.it> wrote:
> What i wish to explore is when a solution is guaranteed to exists
> and i will not loop forever.

In a certain sense, sequences where no solution ever exists are
"special", having density 0.  Of course there are still infinitely
many such sequences.  Uncountably infinite, even.

The only way to determine whether a sequence allows solutions is to
test all its terms.  If the sequence is generated by some finite
algorithm, it may be possible to deduce from a description of the
algorithm (or it might not).


-- 
Tim