Modeling Approaches and Solution Methods for the Lot-Sizing and Raw Material Procurement Problem

• Ano de defesa: 2024
• Autor(a) principal: Tomazella, Caio Paziani
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Dimensionamento de lotes; Lot-sizing; MIP- heuristic; MIP-heurística; Mixed-integer programming; Programação estocástica; Programação inteira-mista; Seleção de fornecedores; Stochastic programming; Supplier selection
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/55/55134/tde-28082024-090710/

This thesis studies the integration of two problems from the supply chain management area: production and raw material procurement. Integrated problems have been gaining increased attention in the literature, due to its relevance in practical applications and to the fact that it results in more cost-efficient solutions. Production planning is an extensively studied topic in the optimisation literature, mostly in the form of lot-sizing problems, while raw material procurement is often related with supplier selection problems. Integrating both problems results in a production plan that takes into consideration not only productive restrictions, but raw material availability, costs and supplier policies such as discounts. This thesis will explore several aspects of the integrated problem, referred to as Integrated Procurement and Lot-Sizing Problem (IPLSP), and its variants. These variants include the IPLSP with perishable inventory, the service-level-driven IPLSP, and the IPLSP with uncertain demand. For the first variant, several MIP-heuristics were proposed to solve the problem, and the best results were obtained with methods that used adaptive operators to improve the local search procedure. The second variant addressed the IPLSP using a novel approach in which, instead of costs, service-levels were optimised. An extensive experimentation showed how this approach differs from the traditional (minimal cost) one, how a more equitable demand fulfilment can be achieved, and how it affects the solution in terms of costs, production and procurement strategies. Incorporating demand uncertainty to the IPLSP increased the complexity of the models, which required an Adjustable Sample Average Approximation heuristic to solve them. Three models with different production flexibility degrees (which decisions can be made after demand realisation) were proposed, and the solutions obtained with the proposed heuristic allowed the calculation of the value of the stochastic solutions and the value of production flexibility. When product perishability was incorporated into the stochastic IPLSP, the solutions indicated that, in some scenarios, the solution included the intentional loss of product inventory in order to reduce costs. In order to avoid these losses, recourse policies were proposed for the addressed models. Overall, this thesis brings contributions to the IPLSP literature in the form of the proposal of novel models, solution methods and meaningful managerial insight for its applications.