Studies Based On PSO On The Estimation In Computer Chinese Chess

Computer game is an important region in artificial intelligence, the effect of the computer game on the artificial intelligence is as same as the drosophila on the genetics. The production in computer game had applied in some practical region,such as, weather forecast, economic control, resource exploration and so on. Although the development of computer Chinese-chess is later than chess, the speed of it is so fast and some excellent Chinese-chess software have been developed, like the ELP, the NEUCHESS, the ELEPHANT EYE and so on. These Chinese-chess software largely depend on a great deal of experience of Chinese-chess masters and lack of the ability to improve themselves strength by self-learning. PSO is introduced into Chinese-chess program in this thesis, and we provide Chinese-chess program with the ability of self-learning. The major works in this paper include three sections:1,The section introduces the main parts of computer Chinese-chess program, including opening-book, endgame databases, move generation, searching algorithm and evaluation function. Some searching algorithms frequently used in computer Chinese-chess and traditional evaluation function are described in detail in this section.2,The basic theory of the particle swarm optimization is introduced, and the pros and cons of the algorithm are discussed.3,Particle swarm optimization (PSO) is improved according to characteristics of Chinese-chess, and the improved PSO is applied to make the Chinese-chess program possess the ability of self-learning. The result of experience has proved the method using PSO to optimize the evaluation function is effectively to enhance the strength of the Chinese-chess program by self-learning.