سال انتشار: ۱۳۸۵

محل انتشار: دوازدهمین کنفرانس سالانه انجمن کامپیوتر ایران

تعداد صفحات: ۴

نویسنده(ها):

Amin Nikanjam – Department of Computer Engineering, Iran University of Science and Technology, Tehran, Iran
Adel Rahmani – Department of Computer Engineering, Iran University of Science and Technology, Tehran, Iran

چکیده:

The Anticipatory Classifier System (ACS) employs the learning classifier system framework and the learning theory of anticipation behavioral control. The resulting evolutionary system can build an internal environmental model and then applies reinforcement learning techniques to develop an optimal set of classifiers. XCSF is another novel version of learning classifier systems (LCS) which introduced the concept of computable classifier prediction and successfully applied to function-approximation problems. In this paper, we apply ACS to function approximation. ACSF is a new version of ACS introduced to develop more accurate approximations.