Robustness of maintenance decisions: uncertainty modelling and value of information

Athena Zitrou, Tim Bedford, Ali Daneshkhah

Research output: Chapter in Book/Report/Conference proceedingConference contribution book

Abstract

In this paper we show how sensitivity analysis for a maintenance optimisation problem can be undertaken by using the concept of Expected Value of Perfect Information (EVPI). This concept is important in a decision-theoretic context such as the maintenance problem, as it allows us to explore the effect of parameter uncertainty on the cost and the resulting recommendations. To reduce the computational effort required for the calculation of EVPIs, we have used Gaussian Process (GP) emulators to approximate the cost rate model. Results from the analysis allow us to identify the most important parameters in terms of the benefit of ‘learning’ by focussing on the partial Expected Value of Perfect Information for a parameter. Assuming that a parameter can become completely known before a maintenance decision is made, the analysis determines the optimal decision and the expected related cost, for the different values of the parameter. This type of analysis can be used to ensure that both maintenance calculations and resulting recommendations are sufficiently robust.
Original languageEnglish
Title of host publicationAdvances in Safety, Reliability and Risk Management
EditorsChristophe Bérengeur, Antoine Grall, Carlos Guedes Soares
Pages940-948
Publication statusPublished - Aug 2011
EventESREL 2011 - 20th European Safety and Reliability Conference - Troyes, France
Duration: 18 Sep 201122 Sep 2011

Conference

ConferenceESREL 2011 - 20th European Safety and Reliability Conference
CountryFrance
CityTroyes
Period18/09/1122/09/11

Keywords

  • maintenance decisions
  • sensitivity analysis
  • Expected Value of Perfect Information
  • EVPI
  • Gaussian process

Fingerprint Dive into the research topics of 'Robustness of maintenance decisions: uncertainty modelling and value of information'. Together they form a unique fingerprint.

Cite this