Strategies for the Automatic Construction of Opening Books
An opening book is an important feature of any game-playing computer program. These books used to be constructed manually by an expert, by storing good moves suggested by theory, or simply by listing all games ever played by strong players [2,5,8]. Interest has recently shifted to automatic opening book construction where positions are selected by a best-first strategy, evaluated using a brute force search and then added to the opening book .
This paper presents the new “drop-out expansion” strategy for automatic opening book construction. It generalizes the previously used best-first strategy and reduces the opportunities for the opponent to force the player out of the book. The algorithm was used to calculate opening books for several games, including Awari and Othello, and helped to win the Awari tournament of the Computer Olympiad .
Keywordsopening book construction expansion strategy best-first Awari Othello
Unable to display preview. Download preview PDF.
- 1.L.V. Allis. Searching for Solutions in Games and Artificial Intelligence. PhD thesis, University of Limburg, Maastricht, The Netherlands, 1994.Google Scholar
- 2.M. G. Brockington. KEYANO Unplugged-The Construction of an Othello Program. Technical Report 97-05, Department of Computing Science, University of Alberta.Google Scholar
- 3.M. Buro. Toward Opening Book Learning. ICCA Journal, 22(2):98–102, 1999.Google Scholar
- 4.R. U. Gasser. Harnessing Computational Resources for Efficient Exhaustive Search. PhD thesis, ETH Zürich, 1995.Google Scholar
- 5.R. M. Hyatt. Book Learning-A Methodology to Tune an Opening Book Automatically. ICCA Journal, 22(1):3–12, 1999.Google Scholar
- 8.J. Schaeffer, R. Lake, P. Lu, and M. Bryant. Chinook: The World Man-Machine Checkers Champion. AI Magazine, 17(1):21–29, 1996.Google Scholar
- 9.K. Thompson. Retrograde Analysis of Certain Endgames. ICCA Journal, 9(3):131–139, 1986.Google Scholar
- 10.R. van der Goot and T. R. Lincke. MarvinWinsAwari Tournament. ICGA Journal, 23(3):173–174, 2000.Google Scholar