How A Computer Chess Program Is Designed?

Since the invention of chess, programmers have made great strides in developing chess techniques and strategies. The idea of coming up with a computer chess program has attracted many people and it began around 1769 with the invention of a chess-playing automaton known as Turk. It was, however, a bogus one.

Genuine machines playing chess have been developed after the introduction of digital computers in the 1950s. From them on, there was a high level of intensity and seriousness in these chess playing machines. The most popular chess computer in history is Deep Blue, which is recognized as the first computer to defeat Garry Kasparov in 1996.

When developing a computer chess program, there are three main issues that needs to be considered namely board representation, search techniques, and evaluation function to determine succeeding moves. Let us briefly explain each issue.

Board representation involves the positioning of the pieces on the board with each position determining a state of the game. They make up the nodes of the search tree of the program. Thus, the structure of the data used in designing the state has an impact on the performance of the search. But, the performance of the chess program depends on the chosen search algorithm and evaluation function.

Search algorithms should not only evaluate the succeeding moves of the program itself, but should likewise look into the counter-moves of the opponents as well. The search tree, which consists of moves, countermoves, and counters to the countermoves, should proceed until a final leaf has been determined. However, it is rarely reached in chess, and the search is restricted for some depth. Likewise, there might be some limitations during the search.

Succeeding moves should be chosen from the board positions developed during the search process. The evaluation function examines the final position. With an increase in the performance of the evaluation function, the chances of winning a game is likewise enhanced.

Some computer chess programmers utilize various search and evaluation functions in different stages of the game. There are three sections in a game of chess namely opening, midgame, and endgame. The middle game employs a standard search and evaluation algorithm. In the case of the opening and endgame, opening and endgame databases compiled from many popular games in chess are used. While this method can save time during these stages, it likewise has the tendency to slow down the performance strengths of computer chess if improperly used.

Archive


Meta

Categories