Move Ordering using Neural Networks

L. Kocsis; J. Uiterwijk; J. Van Den Herik

Move Ordering using Neural Networks

L. Kocsis, J. Uiterwijk, J. Van Den Herik

Networks and Strategic Optimization

Research output: Chapter in Book/Report/Conference proceeding › Chapter › Academic

Abstract

© Springer-Verlag Berlin Heidelberg 2001. The efficiency of alpha-beta search algorithms heavily depends on the order in which the moves are examined. This paper focuses on using neural networks to estimate the likelihood of a move being the best in a certain position. The moves considered more likely to be the best are examined first. We selected Lines of Action as a testing ground. We investigated several schemes to encode the moves in a neural net- work. In the experiments, the best performance was obtained by using one output unit for each possible move of the game. The results indicate that our move-ordering approach can speed up the search with 20 to 50 percent compared with one of the best current alternatives, the history heuristic.

Original language	English
Title of host publication	Proceedings of the 14th International conference on Industrial and engineering applications of artificial intelligence and expert systems: engineering of intelligent systems
Publisher	Springer Verlag
Pages	45-50
Number of pages	6
Volume	2070
ISBN (Print)	9783540422198
Publication status	Published - 2001

Access to Document

http://www.mendeley.com/research/move-ordering-using-neural-networks

Cite this

Kocsis, L., Uiterwijk, J., & Van Den Herik, J. (2001). Move Ordering using Neural Networks. In Proceedings of the 14th International conference on Industrial and engineering applications of artificial intelligence and expert systems: engineering of intelligent systems (Vol. 2070, pp. 45-50). Springer Verlag. http://www.mendeley.com/research/move-ordering-using-neural-networks

@inbook{f132357f19cd4c9ba68cf023aad626b0,

title = "Move Ordering using Neural Networks",

abstract = "{\textcopyright} Springer-Verlag Berlin Heidelberg 2001. The efficiency of alpha-beta search algorithms heavily depends on the order in which the moves are examined. This paper focuses on using neural networks to estimate the likelihood of a move being the best in a certain position. The moves considered more likely to be the best are examined first. We selected Lines of Action as a testing ground. We investigated several schemes to encode the moves in a neural net- work. In the experiments, the best performance was obtained by using one output unit for each possible move of the game. The results indicate that our move-ordering approach can speed up the search with 20 to 50 percent compared with one of the best current alternatives, the history heuristic.",

author = "L. Kocsis and J. Uiterwijk and {Van Den Herik}, J.",

year = "2001",

language = "English",

isbn = "9783540422198",

volume = "2070",

pages = "45--50",

booktitle = "Proceedings of the 14th International conference on Industrial and engineering applications of artificial intelligence and expert systems: engineering of intelligent systems",

publisher = "Springer Verlag",

address = "Germany",

}

Kocsis, L, Uiterwijk, J & Van Den Herik, J 2001, Move Ordering using Neural Networks. in Proceedings of the 14th International conference on Industrial and engineering applications of artificial intelligence and expert systems: engineering of intelligent systems. vol. 2070, Springer Verlag, pp. 45-50. <http://www.mendeley.com/research/move-ordering-using-neural-networks>

Move Ordering using Neural Networks. / Kocsis, L.; Uiterwijk, J.; Van Den Herik, J.

Proceedings of the 14th International conference on Industrial and engineering applications of artificial intelligence and expert systems: engineering of intelligent systems. Vol. 2070 Springer Verlag, 2001. p. 45-50.

Research output: Chapter in Book/Report/Conference proceeding › Chapter › Academic

TY - CHAP

T1 - Move Ordering using Neural Networks

AU - Kocsis, L.

AU - Uiterwijk, J.

AU - Van Den Herik, J.

PY - 2001

Y1 - 2001

N2 - © Springer-Verlag Berlin Heidelberg 2001. The efficiency of alpha-beta search algorithms heavily depends on the order in which the moves are examined. This paper focuses on using neural networks to estimate the likelihood of a move being the best in a certain position. The moves considered more likely to be the best are examined first. We selected Lines of Action as a testing ground. We investigated several schemes to encode the moves in a neural net- work. In the experiments, the best performance was obtained by using one output unit for each possible move of the game. The results indicate that our move-ordering approach can speed up the search with 20 to 50 percent compared with one of the best current alternatives, the history heuristic.

AB - © Springer-Verlag Berlin Heidelberg 2001. The efficiency of alpha-beta search algorithms heavily depends on the order in which the moves are examined. This paper focuses on using neural networks to estimate the likelihood of a move being the best in a certain position. The moves considered more likely to be the best are examined first. We selected Lines of Action as a testing ground. We investigated several schemes to encode the moves in a neural net- work. In the experiments, the best performance was obtained by using one output unit for each possible move of the game. The results indicate that our move-ordering approach can speed up the search with 20 to 50 percent compared with one of the best current alternatives, the history heuristic.

M3 - Chapter

SN - 9783540422198

VL - 2070

SP - 45

EP - 50

BT - Proceedings of the 14th International conference on Industrial and engineering applications of artificial intelligence and expert systems: engineering of intelligent systems

PB - Springer Verlag

ER -