Re: A computer maze

Christian Reynold Byron (crbyron@nospam.socs.uts.edu.au)
Wed, 22 Mar 1995 12:41:25 +1000 (EST)

On Thu, 16 Mar 1995, Farzad Doraius Battiwalla wrote:

> To whom it may concern....
>
> I'm currently interested in writting an AI (Artificially Intelligent)
> maze solving program.
>
> It is known that the computer CANNOT see the full maze and only knows the
> path through which it has already traversed. An analogy could be:
> you have been placed inside a maze, and must find your way through the
> maze, without prior knowledge of the full maze. (Bummer Huh)
>
The only AI I see needed is to "remember" paths already
transversed so that the solver of the maze doesn't retrace paths.
This might occur at T intersections where one branch is a dead
end. Also you need to establish rules for how far down a path a
solver can "see". Once these are established you can then set what
I guess are AI rules for evaluating each path at each intersection.

This still doesn't answer the most important question:-

Why do want to know - What are you up to Fuzz?

Christian
no .sig