THE MOVE-DECISION STRATEGY OF Indigo

Article type: Research Article

Authors: Bouzy, Bruno^{; 1}

Affiliations: Paris, France

Note: [1] C.R.I.P.5, UFR de mathématiques et d’informatique, Université Paris 5, 45 me des Saints-Pères 75270 Paris Cedex 06 France. Email: bouzy@math-info.univ-paris5.fr.

Abstract: This paper describes the move-decision strategy of the Go program INDIGO. It shows that the move-decision process of a Go program can be very different from the processes in other games with a lower complexity than Go. Even though the basic modules are conventional (move generator, evaluation function, and tree search), Indigo uses them in a specific way, viz. adapted to computer Go. The strategy may be of interest to researchers on other mind games that are as complex as Go. The evaluation function can be "fast", "slow" or "strategic". It may include local tree search. The move generation brings about different kinds of moves: "urgent" moves, "life-and-death" moves and "calm" moves. Urgent moves are statically qualified with a global urgency. A two-player quiescence search verifies that the urgent move does not decrease the position evaluation. Calm moves are used within two-player selective global search at a very low depth. Besides, INDIGO also uses single-agent search to refine the strategic importance of the goals. Lastly, INDIGO chooses one out of three (the calm move, the life-and-death move, and the urgent move) to be the global move.

DOI: 10.3233/ICG-2003-26105

Journal: ICGA Journal, vol. 26, no. 1, pp. 14-27, 2003