Dynamic programming for deterministic discrete-time systems with uncertain gain

被引：21

作者：

de Cooman, G ^{[1
]}

Troffaes, MCM ^{[1
]}

机构：

[1] Univ Ghent, Onderzoeksgrp SYSTeMS, B-9052 Zwijnaarde, Belgium

来源：

INTERNATIONAL JOURNAL OF APPROXIMATE REASONING | 2005年 / 39卷 / 2-3期

关键词：

optimal control; dynamic programming; uncertainty; imprecise probabilities; lower previsions; sets of probabilities;

D O I：

10.1016/j.ijar.2004.10.004

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

We generalise the optimisation technique of dynamic programming for discrete-time systems with an uncertain gain function. We assume that uncertainty about the gain function is described by an imprecise probability model, which generalises the well-known Bayesian, or precise, models. We compare various optimality criteria that can be associated with such a model, and which coincide in the precise case: maximality, robust optimality and maximinity. We show that (only) for the first two an optimal feedback can be constructed by solving a Bellman-like equation. (c) 2004 Published by Elsevier Inc.

引用

页码：257 / 278

页数：22

共 20 条

[1]

[Anonymous], 1984, ROBUSTNESS BAYESIAN

[2]

Bellman R., 1957, DYNAMIC PROGRAMMING

[3]

Cheve Morgane., 2000, RISK DECISION POLICY, V5, P151, DOI DOI 10.1017/S1357530900000120

[4]

Couso I, 2000, RISK DECISION POLICY, V5, P165, DOI DOI 10.1017/S1357530900000156

[5]

COZMAN FG, 2003, ISIPTA 03 P 3 INT S, P117

[6] Coherent lower previsions in systems modelling: products and aggregation rules [J].

de Cooman, G ;

Troffaes, MCM .

RELIABILITY ENGINEERING & SYSTEM SAFETY, 2004, 85 (1-3) :113-134

[7]

de Finetti B., 1974, THEORY PROBABILITY C

[8]

DECOOMAN G, 2004, P 10 INT C IPMU 2004, V1, P451

[9]

Giron F.J., 1980, BAYESIAN STAT, P17

[10] Bounded-parameter markov decision processes [J].

Givan, R ;

Leach, S ;

Dean, T .

ARTIFICIAL INTELLIGENCE, 2000, 122 (1-2) :71-109

← 1 2 →