Application of the killer-tree heuristic and the lambda-search method to lines of action

Authors:
Makoto Sakuta;Tsuyoshi Hashimoto;Jun Nagashima;Jos W. H. M. Uiterwijk;Hiroyuki Iida
Affiliations:
Department of Computer Science, Shizuoka University, 3-5-1 Johoku, Hamamatsu, 432-8011 Japan;Department of Computer Science, Shizuoka University, 3-5-1 Johoku, Hamamatsu, 432-8011 Japan;Department of Computer Science, Shizuoka University, 3-5-1 Johoku, Hamamatsu, 432-8011 Japan;Department of Computer Science, Universiteit Maastricht, P.O. Box 616, 6200 MD Maastricht, The Netherlands;Department of Computer Science, Shizuoka University, 3-5-1 Johoku, Hamamatsu, 432-8011 Japan
Venue:
Information Sciences—Informatics and Computer Science: An International Journal - Special issue: Heuristic search and computer game playing III
Year:
2003

Citing 3
Cited 1

The PN -search algorithm: application to tsume-shogi

Artificial Intelligence - Special issue on heuristic search in artificial intelligence
Computer Shogi

Artificial Intelligence - Chips challenging champions: games, computers and Artificial Intelligence
The principal continuation and the killer heuristic

ACM '77 Proceedings of the 1977 annual conference

Evaluation-function based proof-number search

CG'10 Proceedings of the 7th international conference on Computers and games

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper investigates the application of the the killer-tree heuristic (KTH) and the λ1-search method to the endgame of lines of action. Both techniques are developed to be used in the endgame of shogi. They are incorporated into two depth-first searches: iterative deepening αβ and PDS. λ1-Search, in which the move generation is limited to λ1- moves, shows a poor ability to find winning solutions. When using λ1-sorting, in which λ1-moves are put at the front of the move list, the number of nodes searched is reduced considerably, but the solving ability is diminished because of the inefficiency of the move generation. Using the KTH, the number of nodes searched and the solving time are reduced by 5% in iterative deepening αβ and by 20% in PDS. Consequently, the solving ability has been enhanced.