Optimal Strategy for Concurrent Variable Interval Reinforcement Schedule

作     者：Zhenbo Cheng~(1,2),Ming Liang~2,Zhidong Deng~2 1.Information and Engineering College,Zhejiang University of Technology,Hangzhou,310014,China 2.State Key Laboratory on Intelligent Technology and Systems, Tsinghua National Laboratory for Information Science and Technology, Department of Computer Science and Technology,Tsinghua University,Beijing,100084,China 

会议名称：《2010 Chinese Control and Decision Conference》

会议日期：2010年

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 07[理学] 070105[理学-运筹学与控制论] 0701[理学-数学] 

基　　金：supported by National Nature Science Foundation of China under Grant No.60621062 and No.60775040 

关 键 词：Reinforcement Schedule Matching Law Optimal Strategy Matching Strategy 

摘      要：正Herrnstein experimentally studied the choice behavior of pigeons on a special reinforcement schedule,the concurrent variable interval（CVI） schedule,and found a famous matching *** empirical behavior law is remarkably conserved across many kinds of species,but it has been viewed as an irrational behavior,which means that the matching behavior does not maximize *** this paper,we succinctly demonstrate that any strategies leading to matching law can obtain maximal rewards for the CVI reinforcement schedule in discrete time *** addition,we put forward a novel strategy algorithm that can earn the maximal reward in the CVI reinforcement *** results reveal that the matching behavior can be seen as a rational behavior in the reinforcement schedule.