Censored lifetime learning: Optimal Bayesian age-replacement policies

Collin Drent; Stella Kapodistria; Onno J. Boxma

doi:10.1016/j.orl.2020.10.005

Censored lifetime learning: Optimal Bayesian age-replacement policies

Collin Drent (Corresponding author), Stella Kapodistria, Onno J. Boxma

Research output: Contribution to journal › Article › Academic › peer-review

11 Citations (Scopus)

Abstract

We consider a sequence of age-replacement problems with a general lifetime distribution parametrized by an a-priori unknown parameter. There is a trade-off: Preventive replacements are censored but cheap, whereas corrective replacements are uncensored but costly observations of the lifetime distribution. We first analyze the optimal policy for a finite sequence and establish some properties. We then propose a myopic Bayesian policy that almost surely learns the unknown parameter and converges to the optimal policy with full knowledge of the parameter.

Original language	English
Pages (from-to)	827-834
Number of pages	8
Journal	Operations Research Letters
Volume	48
Issue number	6
DOIs	https://doi.org/10.1016/j.orl.2020.10.005
Publication status	Published - Nov 2020

Funding

The research of Collin Drent is supported by the Data Science Flagship framework , a cooperation between the Eindhoven University of Technology and Philips. The research of Stella Kapodistria and Onno Boxma is supported by the NWO Gravitation Programme NETWORKS (Grant No. 024.002.003 ). The authors thank the area editor and the reviewer for their helpful comments and suggestions. The first author thanks Melvin Drent for fruitful discussions in early stages of the preparation of this work.

Keywords

Age-replacement
Asymptotic optimality
Bayesian learning
Censoring
Maintenance

Access to Document

10.1016/j.orl.2020.10.005Licence: CC BY

Cite this

@article{5121c5251b974e69ab4b4ab053d90a2d,

title = "Censored lifetime learning: Optimal Bayesian age-replacement policies",

abstract = "We consider a sequence of age-replacement problems with a general lifetime distribution parametrized by an a-priori unknown parameter. There is a trade-off: Preventive replacements are censored but cheap, whereas corrective replacements are uncensored but costly observations of the lifetime distribution. We first analyze the optimal policy for a finite sequence and establish some properties. We then propose a myopic Bayesian policy that almost surely learns the unknown parameter and converges to the optimal policy with full knowledge of the parameter.",

keywords = "Age-replacement, Asymptotic optimality, Bayesian learning, Censoring, Maintenance",

author = "Collin Drent and Stella Kapodistria and Boxma, {Onno J.}",

year = "2020",

month = nov,

doi = "10.1016/j.orl.2020.10.005",

language = "English",

volume = "48",

pages = "827--834",

journal = "Operations Research Letters",

issn = "0167-6377",

publisher = "Elsevier B.V.",

number = "6",

}

TY - JOUR

T1 - Censored lifetime learning

T2 - Optimal Bayesian age-replacement policies

AU - Drent, Collin

AU - Kapodistria, Stella

AU - Boxma, Onno J.

PY - 2020/11

Y1 - 2020/11

N2 - We consider a sequence of age-replacement problems with a general lifetime distribution parametrized by an a-priori unknown parameter. There is a trade-off: Preventive replacements are censored but cheap, whereas corrective replacements are uncensored but costly observations of the lifetime distribution. We first analyze the optimal policy for a finite sequence and establish some properties. We then propose a myopic Bayesian policy that almost surely learns the unknown parameter and converges to the optimal policy with full knowledge of the parameter.

AB - We consider a sequence of age-replacement problems with a general lifetime distribution parametrized by an a-priori unknown parameter. There is a trade-off: Preventive replacements are censored but cheap, whereas corrective replacements are uncensored but costly observations of the lifetime distribution. We first analyze the optimal policy for a finite sequence and establish some properties. We then propose a myopic Bayesian policy that almost surely learns the unknown parameter and converges to the optimal policy with full knowledge of the parameter.

KW - Age-replacement

KW - Asymptotic optimality

KW - Bayesian learning

KW - Censoring

KW - Maintenance

UR - http://www.scopus.com/inward/record.url?scp=85093956104&partnerID=8YFLogxK

U2 - 10.1016/j.orl.2020.10.005

DO - 10.1016/j.orl.2020.10.005

M3 - Article

AN - SCOPUS:85093956104

SN - 0167-6377

VL - 48

SP - 827

EP - 834

JO - Operations Research Letters

JF - Operations Research Letters

IS - 6

ER -

Censored lifetime learning: Optimal Bayesian age-replacement policies

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this