A finite-horizon, stochastic optimal control policy for a production-inventory system with backlog-dependent lost sales

Gonen Singer, Eugene Khmelnitsky

Research output: Contribution to journalArticlepeer-review

Abstract

This article considers a problem of optimal production control of a one-product-type production-inventory system. The demand is a discrete-time stochastic process, while production of a limited capacity is continious in time. The dynamics of the inventory over a finite time horizon are investigated and an optimal feedback strategy for production control under a backlog-dependent lost sales effect is developed. This strategy is called a target level strategy since, contrary to the existing production-inventory control strategies, it determines an optimal target inventory level as a function of time period and the initial stock at that period. Numerical examples illustrate the applicability of the developed solutions.

Original languageEnglish
Pages (from-to)855-864
Number of pages10
JournalIIE Transactions (Institute of Industrial Engineers)
Volume42
Issue number12
DOIs
StatePublished - Dec 2010

Keywords

  • Inventory control
  • feedback optimal control
  • lost sales
  • perturbation analysis
  • state-co-state dynamics
  • stochastic modeling

Fingerprint

Dive into the research topics of 'A finite-horizon, stochastic optimal control policy for a production-inventory system with backlog-dependent lost sales'. Together they form a unique fingerprint.

Cite this