From: <HIDDEN>
Newsgroups: comp.soft-sys.matlab
Subject: Re: Random sampling with constraints
Date: Mon, 16 Apr 2012 03:39:07 +0000 (UTC)
Organization: The MathWorks, Inc.
Lines: 6
Message-ID: <jmg48r$ffs$>
References: <jmde4t$hd4$> <jmevbj$o98$> <jmg1b2$4gd$>
Reply-To: <HIDDEN>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
X-Trace: 1334547547 15868 (16 Apr 2012 03:39:07 GMT)
NNTP-Posting-Date: Mon, 16 Apr 2012 03:39:07 +0000 (UTC)
X-Newsreader: MATLAB Central Newsreader 1187260
Xref: comp.soft-sys.matlab:764627

"Siva " <> wrote in message <jmg1b2$4gd$>...
> To give you a sense of the scale of the problem, the vector P is [1 x 100000], S is [ 1 x 100], and d is about 50.
- - - - - - - -
  With the length of P you quote, the number of possible selections for S will certainly be too large for the method I mentioned.  On the other hand, that large size makes the idea Bruno mentioned far more effective than it would be for small sized P.  I think you can achieve a reasonably random selection of S using his heuristic method.

Roger Stafford