Optimization of Warehouse Processes Using Machine Learning and Linear Programming
Anastasia Kozlova, Sergei Kurashkin, A A Boyko
Abstract
This paper explores the design of an inventory management model for small warehouses using machine learning and optimization methods. The objective is to develop a system that ensures accurate demand forecasting and automatic calculation of optimal inventory levels, minimizing storage costs and preventing stock shortages. The model incorporates demand forecasting algorithms such as ARIMA, LSTM, and XGBoost, along with inventory optimization methods based on linear programming and genetic algorithms. Special attention is given to integrating an automated inventory replenishment system, as well as monitoring and reporting tools for assessing management efficiency. The proposed model enables companies to reduce operational costs, improve the accuracy and speed of inventory management decision-making, and enhance customer service levels. This work aims to improve inventory management under constrained resources and dynamic market conditions typical of small warehouses.