EURO 2025 Leeds
Abstract Submission

1362. Column Generation Enhanced with Machine Learning: Frameworks for Simultaneous Lot-Sizing and Scheduling

Invited abstract in session WC-15: Topics in Combinatorial Optimization 1, stream Combinatorial Optimization.

Wednesday, 12:30-14:00
Room: Esther Simpson 1.08

Authors (first author is the speaker)

1. Gorkem Yilmaz
Industrial Engineering, Izmir University of Economics
2. Cevdet Utku Şafak
Industrial Engineering, Ozyegin University
3. Erinc Albey
Industrial Engineering, Ozyegin University

Abstract

Capacitated lot-sizing and scheduling problems (CLSP) are essential challenges in production planning, involving optimising both lot sizes and machine scheduling across multiple periods. The complexity of these problems arises from their combinatorial nature and the interaction between lot-sizing and sequence dependent setup times. Traditional optimisation methods struggle with large-scale instances, leading to high computational costs. Column generation has been widely used to address such problems, but the pricing subproblem can become a bottleneck, significantly affecting efficiency. In this paper, we propose a machine learning-based enhancement for the pricing stage of column generation. Specifically, we integrate multi-layer perceptron (MLP) and recurrent neural network (RNN) models to predict promising regions of the solution space, improving the efficiency of column generation by quickly identifying viable solutions. Our method accelerates the optimization process and outperforms the traditional Gurobi solver on the solution quality and the computational time. Experimental results demonstrate that our approach significantly reduces the computational burden, enabling faster and more effective solutions for large-scale CLSP instances. This work highlights the potential of machine learning techniques to improve optimization efficiency in complex, real-world problems, offering a novel approach for solving large-scale capacitated lot-sizing and scheduling problems.

Keywords

Status: accepted

Back to the list of papers