Search and Knowledge in Lines of Action

摘要: This paper describes the design and development of two world-class Lines Action game-playing programs: YL, a three time Computer Olympiad gold-medal winner, Mona, which has dominated international e-mail correspondence play. The underlying philosophy programs is very different: former emphasizes fast efficient search, whereas latter focuses on sophisticated but relatively slow evaluation each board position. In addition to providing technical description program, we explore some long-standing questions trade-offs between search knowledge. These experimental results confirm conclusions made by earlier researchers in domain chess, thus showing that trends are not game-specific. particular, see diminishing returns with additional depth, observe knowledge level program significant impact such experiments.