Newsgroups: comp.ai.genetic
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!Germany.EU.net!EU.net!sun4nl!news.nic.surfnet.nl!ruu.nl!nmr!leef
From: leef@nmr.chem.ruu.nl (Bas Leeflang)
Subject: *** How to solve a complex order problem with GA?????
Message-ID: <1994Dec1.103908.26112@cc.ruu.nl>
Keywords: NMR assignment 
Sender: leef@RUUCI9 (Bas Leeflang)
Organization: ACCU
Date: Thu, 1 Dec 1994 10:39:08 GMT
Lines: 41

Dear Netters,

I am fairly new to the field of genetic algorithms. Therefor I can use some 
advise. 
Most implementations of GA's I have seen thusfar are able to optimise a 
complex function depending on a large number of vaiables, that are 
virtually independent. (At least no mutually excluding)

The problem I am facing is basically to determine the best permutation of a 
large set of object (typically some 100 objects)

The type of information I know is the following :
Information about the RELATIVE ORDER of two objects as elements
i and i+1 and/or i, i+2 and/or i,i+3 and/or i,i+4

Information about the ABSOLUTE POSITION of an object: An object can be found
at location i, whereas it is inpossible (i.e. unlikely) that it is at possition
i+1 for example.

I can come up with an function that produces a 'score' for a certain permutation.

MY PROBLEM is what GA to use to optimise this problem.
Here mutations should be either swaps of two objects or swaps of two ranges of objects.
How is crosover described for a permutation problem?

Are there programs freely available that can handle such a problem?


----------------------------------------------------------------------
                                     Bas R. Leeflang

    _/      _/   _/     _/   _/_/_/  Department of the Bijvoet Center
   _/_/    _/   _/_/ _/_/   _/   _/  for Biomolecular Research
  _/  _/  _/   _/  _/ _/   _/_/_/    Padualaan 8, 3584 CH  Utrecht
 _/    _/_/   _/     _/   _/   _/    Tel. : int+31.30.533295
_/      _/   _/     _/   _/    _/    Fax  : int+31.30.537623
                                     Email: leef@nmr.chem.ruu.nl


-- 
Bas Leeflang
