A and B strategy

from Wikipedia, the free encyclopedia

A and B strategy are terms used in the programming of strategy games , especially computer chess .

description

A program for playing a strategy game usually looks at a search tree that is part of the game tree .

According to Claude Shannon, A-strategy is a method that calculates all possible combinations of moves and counter moves to determine the best move, up to a certain depth (number of consecutive moves), which is limited by computing power and available time. The positions reached are evaluated heuristically , and the moves that have led to them are evaluated according to the minimax principle . The A strategy is also known as the brute force method .

In contrast to the A strategy, a chess program plays according to the B strategy if it only searches plausible - and not all - move orders when analyzing a position. The moves are evaluated heuristically and only those with a high value are included in the search tree. The B strategy is sometimes understood as an attempt to reproduce the human thought process when analyzing variants. Man also does not calculate all legal moves in a position, but recognizes certain features of the position and then selects moves for closer examination that are promising with regard to the positional features.

Because the number of variants grows more slowly with depth than with the A strategy, the variants can be calculated up to a greater depth. This enables the program to identify tactical combinations that are beyond the horizon of the A-strategy. On the other hand, it sometimes overlooks a twist by giving a good move a low rating and thereby excluding it from the search tree from the outset.

The first attempt to write such a chess program was made by Allen Newell , John Shaw and Herbert A. Simon between 1955 and 1958 . It practically failed, and people began to understand that implementing such a program is far more difficult than initially thought. Modern chess programs predominantly use modified forms of the A-strategy, which pursue individual promising or tactically critical variants more deeply than others.

In contrast to computer chess, the brute force methods reach their limits in the board game Go. Research in the direction of intelligent solution approaches (B strategy) is still in full swing, unlike in chess .

literature

  • Edmund Bruns: The game of chess as a phenomenon of the cultural history of the 19th and 20th centuries . LIT Verlag, Münster 2003, ISBN 978-3-8258-6546-7 , pp. 307-308 ( online view ).