Alumni

The Hex Player Project An Experiment in Self Directed Machine Learning

View Full Record
Description: The genetic algorithm is known to obtain solutions for many different problems, including some instances where the algorithm led computers to competently play board games without the benefit of human expertise. This paper reports on the planning and results of an experiment to determine if a genetic algorithm could develop machine intelligence capable of playing the game of Hex without any human knowledge. It is found that, while the algorithm developed an approach to play the game and was able to defeat a player making random moves, the approach did not have the flexibility needed to compete against human players.
Language: English
Format: Degree Work