ADAPTIVE-CONTROL OF MARKOV-CHAINS - FINITE PARAMETER SET

被引：67

作者：

BORKAR, V ^{[1
]}

VARAIYA, P ^{[1
]}

机构：

[1] UNIV CALIF BERKELEY,ELECTR RES LAB,BERKELEY,CA 94720

来源：

IEEE TRANSACTIONS ON AUTOMATIC CONTROL | 1979年 / 24卷 / 06期

关键词：

D O I：

10.1109/TAC.1979.1102191

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

Consider a controlled Markov chain whose transition probabilities depend upon an unknown parameter a taking values in finite set A. To each a is associated a prespecified stationary control law ϕ(ðŗ‚). The adaptive control law selects at each time t the control action indicated by ϕ(ðŗ‚1) where ϕ(ðŗ‚1) is the maximum likelihood estimate of a. It is shown that (ðŗ‚1) converges to a parameter ðŗ‚* such that the “closed-loop” transition probabilities corresponding to a* and ϕ(ðŗ‚*) are the same as those corresponding to ðŗ‚0 and ϕ(ðŗ‚*) where ðŗ‚0 is the true parameter. The situation when ðŗ‚0 does not belong to the model setA is briefly discussed. Copyright © 1979 by The Institute of Electricala and Electronics Engineers Inc.

引用

页码：953 / 957

页数：5

共 8 条

[1] SELF TUNING REGULATORS [J].