Approximations on risk-averse Markov decision processes

We consider the problem of approximating the values and the optimal policies in risk-averse discounted Markov Decision Processes with infinite horizon. We study the properties of the rolling horizon and the approximate rolling horizon procedures, proving bounds which imply the convergence of the pro...

ver descrição completa

Detalhes bibliográficos
Autores: Della Vecchia, Eugenio Martín, Di Marco, Silvia Cristina, Jean Marie, Alain
Formato: artículo
Estado:Versión publicada
Fecha de publicación:2013
País:Argentina
Recursos:Consejo Nacional de Investigaciones Científicas y Técnicas
Repositorio:CONICET Digital (CONICET)
Idioma:inglés
OAI Identifier:oai:ri.conicet.gov.ar:11336/15583
Acesso em linha:http://hdl.handle.net/11336/15583
Access Level:acceso abierto
Palavra-chave:MARKOV DECISION PROCESSES
RISK AVERSION
ROLLING HORIZON
https://purl.org/becyt/ford/1.1
https://purl.org/becyt/ford/1
Descrição
Resumo:We consider the problem of approximating the values and the optimal policies in risk-averse discounted Markov Decision Processes with infinite horizon. We study the properties of the rolling horizon and the approximate rolling horizon procedures, proving bounds which imply the convergence of the procedures when the horizon length tends to infinity. We also analyze the effects of uncertainties on the transition probabilities, the cost functions and the discount factors.