Convergence analysis of an incremental approach to online inverse reinforcement learning

作     者：Zhuo-jun JIN Hui QIAN Shen-yi CHEN Miao-liang ZHU 

作者机构：School of Computer Science and Technology Zhejiang University Hangzhou 310027 China 

出 版 物：《Journal of Zhejiang University-Science C(Computers and Electronics)》 (浙江大学学报C辑（计算机与电子（英文版）)

年 卷 期：2011年第12卷第1期

页      面：17-24页

核心收录：

学科分类：0810[工学-信息与通信工程] 12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 0805[工学-材料科学与工程（可授工学、理学学位）] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：Project (No.90820306) supported by the National Natural Science Foundation of China 

主　　题：Incremental approach Reward recovering Online learning Inverse reinforcement learning Markov decision process 

摘      要：Interest in inverse reinforcement learning (IRL) has recently increased,that is,interest in the problem of recovering the reward function underlying a Markov decision process (MDP) given the dynamics of the system and the behavior of an *** paper deals with an incremental approach to online ***,the convergence property of the incremental method for the IRL problem was investigated,and the bounds of both the mistake number during the learning process and regret were provided by using a detailed *** an online algorithm based on incremental error correcting was derived to deal with the IRL *** key idea is to add an increment to the current reward estimate each time an action mismatch *** leads to an estimate that approaches a target optimal *** proposed method was tested in a driving simulation experiment and found to be able to efficiently recover an adequate reward function.