Fine-Grained Decision-Theoretic Search Control
This work addresses search efficiency for game-playing AI, but it is incremental as it builds on prior decision-theoretic frameworks.
The paper tackles the problem of inefficient search control in decision-theoretic methods by extending analysis to individual successor evaluations, resulting in an improved MGSS* algorithm that shows empirical gains in the game of Othello.
Decision-theoretic control of search has previously used as its basic unit. of computation the generation and evaluation of a complete set of successors. Although this simplifies analysis, it results in some lost opportunities for pruning and satisficing. This paper therefore extends the analysis of the value of computation to cover individual successor evaluations. The analytic techniques used may prove useful for control of reasoning in more general settings. A formula is developed for the expected value of a node, k of whose n successors have been evaluated. This formula is used to estimate the value of expanding further successors, using a general formula for the value of a computation in game-playing developed in earlier work. We exhibit an improved version of the MGSS* algorithm, giving empirical results for the game of Othello.