A Tetris AI with an adventurous strategy based on Attribute Theory Method

During recent years, an amount of researchers are becoming interested in the popular video game Tetris. However, most of the researchers concentrate on eliminating as many rows as possible only, without considerations on how many rows are eliminated within each action. In this paper, we simulate human players´ behavior and try to increase the average rows eliminated each action as far as we can, and during this procedure we basically use Attribute Theory Method. We call it adventurous strategy because it is venturesome in some respect. We also introduce a danger evaluation to counteract the adverse factors of this strategy.