Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Schaeffer, Long-range planning in computer chess, 1983

Author: Steven Edwards

Date: 02:46:05 05/25/05

Go up one level in this thread


On May 25, 2005 at 01:19:41, Joshua Shriver wrote:

>By chance is this article available online? If not I'll check my uni's archives
>for a pulp version.

It's not online as far as I know.  But it's worth a trip to the library.

The key points of Paradise:

1. Use of a programming environment that supports simple and extendible symbolic
representation.

2. Explicit representations for relatively abstract ideas like plans, goals, and
threats.

3. A blackboard subsystem for accessing instances of pattern matching along with
a somewhat classical production system that produces the instances.

4. Encapsulation of knowledge sources as non disjoint sets of production rules.

5. The concept of a plan as being a search control program with plan formation
using more time than the rest of the search process.

6. A causality facility for relatively aggressive forward pruning and a
specialized quiescence search.

7. A whole tree retention vs. traditional single/current path only.

8. Satisfaction with finding a good move vs. finding the best move.

9. Production of a somewhat readable search narrative.



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.