Abordagem heurística da relaxação surrogate para o planejamento da produção de pedidos DE PEDIDOS

• Ano de defesa: 2019
• Autor(a) principal: ALVES, Giovanna Abreu
• Orientador(a): CAMARGO, Victor Claudio Bento de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Triângulo Mineiro
• Programa de Pós-Graduação: Programa de Mestrado Profissional em Inovação Tecnológica
• Departamento: Instituto de Ciências Tecnológicas e Exatas - ICTE::Programa de Mestrado Profissional em Inovação Tecnológica
• País: Brasil
• Palavras-chave em Português: Fundições.; Planejamento de Pedidos.; Dimensionamento de Lotes.; Relaxação Surrogate.
• Palavras-chave em Inglês: Foundries.; Orders Planning.; Lot sizing.; Surrogate Relaxation.
• Área do conhecimento CNPq: Engenharia de Produção; Pesquisa Operacional
• Link de acesso: http://bdtd.uftm.edu.br/handle/tede/760

The production process of a small foundry is characterized by a furnace fed by raw materials, producing metal alloys, with limited capacity to meet the demand for orders. Lot sizing and scheduling problem for foundries management has been rather studied in the literature, but in most of these works the production, storage and delay of the items in a dissociate form is studied, what is proposed in this work is the treatment production planning, that means, consider the order to which an item belongs. An order contains several items that may or may not depend on each other, that is, they can not be delivered separately. In this sense, it is important a good production planning that seeks to meet the highest number of orders without delays, in order to reduce costs. One way of solving the problem of sizing and batch sequencing is to determine a relaxation for the original problem and from it to determine initial solutions for heuristic methods of resolution. In this work we propose a heuristic approach, based on the Relaxation Surrogate for the problem of planning the production of orders in a small foundry. We present theoretical results that confirm that the proposed model is a relaxation of an existing model in the literature. The solutions obtained by computationally solving the relaxed model are used as initial solutions of a heuristic that solves the problem of order planning efficiently.