EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

693. Stochastic reoptimisation model for inventory management

Invited abstract in session WB-28: Advancements of OR-analytics in statistics, machine learning and data science 9, stream Advancements of OR-analytics in statistics, machine learning and data science.

Wednesday, 10:30-12:00
Room: 065 (building: 208)

Authors (first author is the speaker)

1. Zahra Namazian
Optimization, Data Science and AI, Faculty of Information Technology, Faculty of IT, Monash University
2. John Betts
Data Science and AI, Monash University
3. Peter Stuckey
Faculty of Information Technology, Monash University

Abstract

Traditional inventory control approaches typically begin by estimating the demand distribution within a predefined family of distributions based on historical observations.
The traditional inventory models use these estimated distributions to find the optimal order policy. However, these approaches often rely on insufficient information, for example, only the mean and standard deviation of demand, which may not adequately capture demand changes over time. By contrast with these approaches, we utilize a Light GBM model to predict the daily demand. By rebuilding the LightGBM model at each time period we track changes over time. Consequently, we propose an innovative approach that integrates the LightGBM model within a two-stage stochastic optimisation framework executed on a daily basis. In particular, by employing this integrated predictive model, we achieve a more responsive tracking of changes in demand and make the replenishment decision through continuous reoptimisation of stochastic programming on a daily basis. The effectiveness of our approach is demonstrated in our case study, a retail company dataset, Where the new approach results in reducing inventory management costs without significantly affecting customer service levels, distinguishing it from traditional approaches.

Keywords

Status: accepted

Back to the list of papers