Re: computer-go: Graph Theory

[Heikki Levanto <heikki@xxxxxx>]

On Wed, May 01, 2002 at 10:29:49PM -0700, Tyler D Sellon wrote:
> The reason I found this interested is that there is an effecient algorithm
> for finding the cut vertices of a given connected graph.  This allows
> one to ignore specific properties of the shape in favor of more general
> ones.  This allows me to avoid using pattern recognition, which is good
> since I don't know anything about it.

>From what I have seen here, this is indeed a simple way to solve the most
trivial of life and death problems, but they get *much* more complex when
the eye space is not completely enclosed, and when there are enemy stones
inside or around it...

This kind of algorithm may be of some use in evaluating the leaves of a
larger search tree, but only if it is truly fast, and if you have a really
fast way to isolate the eye space in question.  Without fairly deep reading,
it will only solve the most elementary kind of problems, which can easily
be solved by other means as well.

- Heikki

-- 
Heikki Levanto  LSD - Levanto Software Development   <heikki@xxxxxxxxxxxxxxxxx>