Scheduling wine bottling operations with multiple lines and sequence-dependent set-up times: Robust formulation and a decomposition solution approach

Mac Cawley A.; Maturana, S; Pascual, R.; Tortorella, G

Keywords: optimisation, scheduling, Decomposition algorithm, Sequence-Dependent Set-up Times, Robust formulation

Abstract

Planning and scheduling the multiple bottling lines of large wineries can significantly increase or reduce their inventory cost and bottling lines operating costs. It also determines how well they meet their clients’ demand, which is critical for wineries. A good plan must consider demand, labour availability, bottling and labelling lines capacities, and the required materials. It also must ensure that the product is available on time, simultaneously keeping finished product inventory as low as possible, and efficiently using the lines and labour. For this, the planner must consider a large number of parameters, like customer orders, the lines’ production and labour capacities, bottling supplies availability, storage capacity, and labour costs. Furthermore, bottling wine has sequence-dependent set-up times. Changing from bottling red wine to white requires twice the set-up time needed to change from white to red, due to line cleaning. This makes planning and scheduling multiple bottling lines a complex process. We developed two models to support this process for a large winery: a basic one and one that generates more robust plans. Since solving this problem is difficult, we devised a decomposition algorithm that takes advantage of the different set-up types to reduce the time required to solve large problem instances. The results showed cost reductions of approximately 15% to 30% on test problems. We could also generate more robust plans without a significant cost increase. The model and solution approach were used to implement a decision support system for a large winery.

Más información

Título de la Revista: EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
Editorial: Elsevier
Fecha de publicación: 2022
Idioma: Ingles
URL: https://www.sciencedirect.com/science/article/abs/pii/S0377221722001734
DOI:

10.1016/j.ejor.2022.02.054