1. Searching for compound goals using relevancy zones in the game of Go
- Author
-
Jan Ramon, Tom Croonenborghs, VanDenHerik, HJ, Bjornsson, Y, and Netanyahu, NS
- Subjects
Theoretical computer science ,Search algorithm ,Computer science ,business.industry ,Branching factor ,Contrast (statistics) ,Local search (optimization) ,Artificial intelligence ,business - Abstract
In complex games with a high branching factor, global alphabeta search is computationally infeasible. One way to overcome this problem is by using selective goal-directed search algorithms. These goal-directed searches can use relevancy zones to determine which part of the board influences the goal. In this paper, we propose a general method that uses these relevancy zones for searching for compound goals. A compound goal is constructed from less complex atomic goals, using the standard connectives. In contrast to other approaches that treat goals separately in the search phase, compound goal search obtains exact results. ispartof: pages:113-128 ispartof: Lecture Notes in Computer Science vol:3846 pages:113-128 ispartof: 4th International Conference on Computers and Games location:Ramat-Gan, Israel date:5 Jul - 7 Jul 2004 status: published
- Published
- 2006