Developing Artificial Intelligence Algorithm For Playing Chess

OData support
Supervisor:
Dr. Juhász Sándor
Department of Automation and Applied Informatics

Since the starting of artificial intelligence research and the release of the first computers in the 1940s, scientists compared the performance of computers and human minds. The chess game was a great choice for such experiments. For a long time machines could not defeat humans, but in 1997 an IBM computer called Deep Blue - specially developed for that purpose - beat the actual world champion, Garry Kasparov.

The topic of my thesis was to create a chess program, which is a match for an active, competitive chess player as well.

The thesis evaluates the efficiency of the main board representation and move generation techniques and implements them with some changes. I measure the performance of the critical functions. My thesis presents the search algorithms used in artificial intelligence, and compare them. I define the aspects for the machine, which have to be considered in order to calculate the relative value of a particular position.

I have tested my finished program against other computers and online opponents on a chess server as well, and determined the strength of it. The outcome of my work is a chess AI with an ELO rating of about 1750.

Downloads

Please sign in to download the files of this thesis.