Newsgroups: alt.comp.hardware.homebuilt,alt.computer.consultants,comp.ai,comp.periphs,comp.programming,comp.robotics
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!gatech!rutgers!rochester!rit!isc-newsserver!vaxa.isc.rit.edu!RXR8697
From: rxr8697@vaxa.isc.rit.edu
Subject: Re: Help ! maze solving
Message-ID: <1994Nov25.210632.17464@ultb.isc.rit.edu>
Sender: news@ultb.isc.rit.edu (USENET News System)
Nntp-Posting-Host: vaxa.isc.rit.edu
Reply-To: rxr8697@vaxa.isc.rit.edu
Organization: Rochester Institute of Technology
References: <3avo57$kut@news.kth.se> <CztFEo.9u1@cruzio.com>,<allenbach.16.2ED5C439@di.epfl.ch>
Date: Fri, 25 Nov 1994 21:06:32 GMT
Lines: 38
Xref: glinda.oz.cs.cmu.edu comp.ai:25498 comp.periphs:9994 comp.programming:13059 comp.robotics:15672

In article <allenbach.16.2ED5C439@di.epfl.ch>, allenbach@di.epfl.ch (Philippe Allenbach) writes:
>
>Jan Soderman writes:
>
>>> Hello !
>>> The problem for us is to find a suitable algorithm or program to searh and
>>solve
>>> the maze. I would be wery glad if anyone has any ideas that might help.
>>> 
>>> 
>>> Jan Soderman.
>
>You may try genetic algorithms, people are doing this, in particular at the 
>Ecole Polytechnique de Grenoble (France)
>Try to ask on comp.ai.genetic
>
>Philippe Allenbach
>allenbach@di.epfl.ch

Genetic algorithms to solve a 16x16 maze! I think this might be too much
overhead for such a simple problem. Use a graph and figure out every possible
path from the start to finish points. Measure the cost of taking the different
routes making sure to give a higher cost to a turn vs a straight line path.
The path with the smallest cost is your solution.

To figure out the different paths start at a point and map every single
possible direction. Eliminate anything that takes you to a dead end and
you will have a number of possible paths.

This kind of stuff should be kept simple to be effective. Also use a high
level programming language to implement the algorithm as this is too tedious
to do in assembly.

Rudi R.

Please email all replies, flames and so on to 'rxr8697@ritvax.isc.rit.edu'.

Rudi R.
