摘要
在游戏开发的人工智能领域中,关键部分是学习过程。而经典的学习系统需要花费很多时间忘记旧规则,同时学习新规则。为解决这个问题,提出使用支持向量机技术,并为这种学习分类器引进了一个时间概念,这样,系统就可以容易地忘记过去的事情,同时文中也为支持向量的数量限制和CPU消耗的限制提出了解决方法。
In the field of the artificial intelligence for game development, the learning process is very impor- tant. In the classical learning process much time is taken to forget old rules learn new rules. In order to solve the problem, this paper puts forward the method of Support Vector Machines. By introducing the time concept for this method, the system can easily forget old rules. This paper also brings forward an effective solution to both the limit of amount of Support Vector Machines and the limit of CPU＇s consumption.
出处
《电子科技》
2012年第9期60-62,66,共4页
Electronic Science and Technology
关键词
游戏开发
人工智能
支持向量机
学习过程
game development
artificial intelligence
support vector machines
learning process
