| 摘要: |
| 通用对弈游戏（general game playing，简称GGP）是致力于提高机器的通用游戏智能的研究领域.与专用游戏智能程序不同，GGP玩家直到游戏开始时才获得游戏规则，从而避免依赖于人类关于特定游戏的经验.GGP研究发展至今，已在游戏表示、搜索算法、状态估值等方面做了深入探索，并在知识迁移等方面做出了尝试.GGP研究的进展在一定程度上代表了通用人工智能的发展，因而是值得关注的. |
| 关键词: 通用 游戏 逻辑 人工智能 知识表示 |
| DOI：10.13328/j.cnki.jos.004928 |
| 分类号: |
| 基金项目: |
|
| General Game Playing:A Research Field for Exploring Machine Intelligence in Games |
|
ZHANG Hai-Feng, LIU Dang-Yi, LI Wen-Xin
|
|
School of Electronics Engineering and Computer Science, Peking University, Beijing 100871, China
|
| Abstract: |
| General game playing (GGP) is a research field for improving the general gaming intelligence of machines. It is different from specific game playing in that GGP players do not know game rules before the game begins, which makes them independent from human experience on specific games. Until now, GGP researchers have deeply explored many problems, such as game representation, search algorithm, and state evaluation. Also, some efforts have been made on knowledge transfer. The progress of GGP research represents the development of artificial general intelligence to some extent, which makes it remarkable. |
| Key words: general game logic artificial intelligence knowledge representation |
