The Cost of Risk-Aversion In Inventory Management: An (s,S) Case Study

Ebru Angun, Jack Kleijnen

Research output: Working paperDiscussion paperOther research output

130 Downloads (Pure)

Abstract

To model a risk-averse attitude (instead of a risk-neutral attitude), we may require that the 90% quantile (instead of the expected value) of a specific uncertain (or random) response satisfy a prespecified threshold which corresponds with a chance constraint. We include a case study; namely, an (s, S) inventory model that is specified in the literature. In this study we require that the 90% quantile of the service level exceed a prespecified threshold. So, we need to estimate the optimal values of s and S, which satisfy this service-level constraint while minimizing the expected inventory cost. To solve the resulting constrained optimization problem, we apply a recent variant of “efficient global optimization” (also known as “Bayesian optimization” and related to “active” machine learning). Our numerical results for the case study imply that the mean inventory cost increases by 2.4% if management is risk-averse instead of risk-neutral.
Original languageEnglish
Place of PublicationTilburg
PublisherCentER, Center for Economic Research
Number of pages21
Volume2024-005
Publication statusPublished - 5 Mar 2024

Publication series

NameCentER Discussion Paper
Volume2024-005

Keywords

  • risk aversion
  • chance constraint
  • inventory management
  • simulation optimization

Fingerprint

Dive into the research topics of 'The Cost of Risk-Aversion In Inventory Management: An (s,S) Case Study'. Together they form a unique fingerprint.

Cite this