home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.theory:2849 sci.crypt:6603 sci.math:17912 rec.puzzles:8312
- Newsgroups: comp.theory,sci.crypt,sci.math,rec.puzzles
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!sdd.hp.com!think.com!linus!philabs!acheron!scifi!watson!yktnews!admin!wo0z!lwloen
- From: lwloen@rchland.vnet.ibm.com (Larry Loen)
- Subject: Re: Looking for random permutation generation algorithms
- Sender: news@rchland.ibm.com
- Message-ID: <1993Jan09.220921.19430@rchland.ibm.com>
- Date: Sat, 09 Jan 1993 22:09:21 GMT
- Reply-To: lwloen@rchland.vnet.ibm.com
- Disclaimer: This posting represents the poster's views, not necessarily those of IBM
- References: <1993Jan6.014749.15323@ee.ubc.ca> <1993Jan6.224804.6116@dsuvax.dsu.edu>
- Nntp-Posting-Host: wo0z.rchland.ibm.com
- Organization: IBM Rochester
- Lines: 29
-
-
- If I understand the request correctly, the original poster is looking
- for a way of generating a pseudo random string of digits including all
- number 0, 1. . . n-1 for n numbers, without repeating any.
-
- The simplest way to do this comes from APL. What you do is generate
- a string of n numbers using any suitable random number generator.
-
- Then, you do a "sort" of the numbers using the "index array" method.
-
- That is, you leave the original numbers in place, and sort an array
- of "indexes" to the original data.
-
- As it happens, that sorted array will, itself, provide the array
- required. In other words, the original random numbers (that may have
- had repeats) is discarded after the sort, because they were merely means
- to an end.
-
- Now, if that doesn't quite "cut" what is looked for, here, such an array
- can probably be used to manipulate some sort of starting permutation
- in a pseudo-random way with only a little more thought. However, since
- the numbers don't repeat, it would seem to form a permutation-substitution
- by simply having index i replaced by index i+1 and so forth.
-
- If one has a built-in sort handy, it is very easy to do.
-
- --
- Larry W. Loen | My Opinions are decidedly my own, so please
- | do not attribute them to my employer
-