Abordagens para resolver o problema integrado de dimensionamento de lotes e scheduling em máquinas paralelas

• Ano de defesa: 2019
• Autor(a) principal: Carvalho, Desiree Maldonado [UNIFESP]
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Paulo (UNIFESP)
• Palavras-chave em Português: Integrated Problem; Paralel Machines; Lotsizing; Scheduling; Problem Integrado; Máquinas Paralelas; Dimensionamento De Lotes; Preparação Não Triangular
• Link de acesso: https://sucupira.capes.gov.br/sucupira/public/consultas/coleta/trabalhoConclusao/viewTrabalhoConclusao.jsf?popup=true&id_trabalho=7716833; https://repositorio.unifesp.br/handle/11600/59480

This study approaches the integrated lot sizing and scheduling problem (ILSSP), in which non-identical machines work in parallel with sequence-dependent non-triangular (NT) setup costs and times, setup carr y-over and capacity limitation. The aim of the studied ILSSP, called here ILSSP-NT on parallel machines, is to deter mine a production planning and tasks sequencing which meet the period demands without delay, in such a way that the total costs of production, machine setup and inventor y are minimized. The dearth of literature on the ILSSP-NT on parallel machines and the researchers interest to approach it motivated the study perfor med in this thesis. The aim of this thesis is the development of efficient solution methods to the ILSSP-NT on parallel machines, in particular, the proposal of matheuristics. To this end, it was performed an analysis of the main for mulations designed to the ILSSP-NT found in the literature, adapted to the problem with parallel machines. The pur pose was defining which for mulation could contribute to the best computational performance of solver CPLEX v. 12. 7. 1 limited by a period of time to solve instances from the literature, adapted to the problem with non-triangular setup costs and times. Besides this analysis, in a first moment, we performed a study on the capacitated lot sizing problem (CLSP) on parallel machines where we proposed a math-heuristic based in a Lagrangian heuristic and a method known as kernel search. Such method aimed the solution of the CLSP on parallel machines with SC and inspired the development of one of the methods proposed to the ILSSP-NT on parallel machines. The methods we propose to find solutions to the ILSSP-NT on parallel machines consist, virtually in the hybridization of relax-and-fix and fix-and-optimize methods with path-relinking and kernel search heuristics. Computational experiments attest the good performance of the proposed methods to find solutions to the CLSP, the ILSSP and the ILSSP-NT on parallel machines.