A policy improvement method in constrained stochastic dynamic programming

被引：9
|
作者：
Chang, Hyeong Soo [1 ]
机构：
[1] Sogang Univ, Dept Comp Sci & Engn, Seoul 121742, South Korea
[2] Sogang Univ, Program Integrated Biotechnol, Seoul 121742, South Korea
来源：
IEEE TRANSACTIONS ON AUTOMATIC CONTROL | 2006年 / 51卷 / 09期
关键词：
constrained Markov decision process; dynamic programming; policy improvement; policy iteration;
D O I：
10.1109/TAC.2006.880801
中图分类号：
TP [自动化技术、计算机技术];
学科分类号：
0812 ;
摘要：
This note presents a formal method of improving a given base-policy such that the performance of the resulting policy is no worse than that of the base-policy at all states in constrained stochastic dynamic programming. We consider finite horizon and discounted infinite horizon cases. The improvement method induces a policy iteration-type algorithm that converges to a local optimal policy.
引用
下载
收藏
页码：1523 / 1526
页数：4
相关论文
共 50 条
12345