Otimização do planejamento tático da cadeia de suprimentos: formulações e métodos

• Ano de defesa: 2015
• Autor(a) principal: Joao Flavio de Freitas Almeida
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Decomposição de Benders; Programação estocástica robusta; Planejamento da cadeia de suprimentos; Cadeia de suprimentos; Logística empresarial; Método de decomposição; Programação estocástica; Engenharia de produção
• Link de acesso: http://hdl.handle.net/1843/BUBD-9WXFCB

Manufacturing industries are characterized by having complex logistics and production systems. In the planning area, goals of multiple business divisions such as marketing, distribution, manufacturing and purchases are often conicting, requiring the development of a unied and rigorous structure capable of capturing the various synergies and tradeos involved. In this thesis, we deal with the optimization of tactical supply chain planning for manufacturing industries. We present multi-product, multi-modal and multi-period formulations to integrate medium-term decisions addressing the supply, production and distribution of a four echelons supply chain: suppliers, factories, distribution centers and customers. We develop deterministic and stochastic formulations addressed by stochastic programming and robust stochastic programming. To solve the large scale stochastic problems, we developed methods of stochastic decomposition based on Benders (1962). The developed models and methods are evaluated by a computational study. The analysis also evaluate the exibility of the supply chain. Such approach has not been explored in the literature, as discussed in the recent study of Esmaeilikia et al. (2014). Finally, a complete case study evaluates the application of stochastic programming to the annual tactical planning of a steel making company supply chain. We describe logistics, operations and the model development. Computational results demonstrate the superior performance of the proposed decomposition method compared to the monolithic formulation. The quality of the stochastic programming solution is also demonstrated.