Replacement optimisation for public infrastructure assets: Quantitative optimisation modelling taking typical public infrastructure related features into account

Research output: ThesisDissertation (TU Delft)

88 Downloads (Pure)

Abstract

Ageing infrastructures and shortage of financing induce the need for optimising public infrastructure replacements. From an economic perspective, classical net present value comparison is traditionally the method of choice to decide on investments and replacements. The current research observes that typical infrastructure related features make the classical net present value comparison less suitable in its application for optimising infrastructure replacements. Especially the low discount rate of public sector organisations, price increases and price uncertainty contribute to this phenomenon in which the application of classical net present value comparison leads to suboptimal timing and costs. This observation led to the development of six dedicated replacement optimisation models for common types of infrastructure replacement challenges. A decision support guideline is provided to assist in selecting an appropriate model based on the sequence of intervention strategies, the development of forecasted cash flows and whether uncertainty is involved. The quantitative replacement optimisation models function as blueprints for similar challenges and support a wider decision-making context.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • Delft University of Technology
Supervisors/Advisors
  • Bakker, H.L.M., Supervisor
  • Kapelan, Z., Supervisor
Award date15 Jul 2020
Print ISBNs978‐94‐028‐1965‐6
DOIs
Publication statusPublished - 25 Mar 2020

Keywords

  • replacement
  • optimisation
  • public infrastructure
  • reliability
  • real options
  • uncertainty
  • Markov decision process

Fingerprint

Dive into the research topics of 'Replacement optimisation for public infrastructure assets: Quantitative optimisation modelling taking typical public infrastructure related features into account'. Together they form a unique fingerprint.

Cite this