New algorithms exploiting randomness in lead times in inventory systems

In this work, we examine a traditional single-item inventory system where unfilled demands result in backlog. The time it takes to restock varies at random and is independent and identically distributed, leading to orders crossing paths.The research conducted by (1) demonstrates that the randomness of lead times in inventory systems can potentially result in infinite improvements compared to constant lead times. The aim of this project is to investigate the feasibility of achieving such improvements while considering practical system constraints and identifying policies that are both effective and practical. This study introduces two new policies, Adaptive and Pipeline, in addition to the discrete-time version of the Generalized Base Stock policy introduced in (1) and Constant Base Stock Policy ((2)). The performance of the four policies are evaluated through simulations, with a focus on the impact of lead time distributions. The results indicate that the proposed policies can lead to significant performance improvements under practical constraints, with larger improvements observed when lead time variability is higher. We find that the Pipeline policy is found to be the most effective and practical for implementation.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Industrial Engineering

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2023

Author dc:creator

• Taneja, Aditi

Contributors dc:contributor

• Stolyar, Aleksandr
• Wang, Qiong

• Gbs
• Pipeline
• Cbs
• Adaptive

Statement dc:rights

• Copyright 2023 Aditi Taneja

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/120578

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F120578