Planejamento da produção de impressoras por dimensionamento de lotes e sequenciamento: um estudo de caso

• Ano de defesa: 2009
• Autor(a) principal: Joao Flavio de Freitas Almeida
• Outros Autores: Samuel Vieira Conceicao
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Sequenciamento; Planejamento da produção; Dimensionamento de lotes; Engenharia de produção
• Link de acesso: http://hdl.handle.net/1843/MCBR-8A9H5W

In this thesis we consider a real problem of lot-sizing and scheduling printers production in a worldwide EMS (Electronic Manufacturing Service) company. The environment is characterized by the production of four different types of printers on two non identical manufacturing lines. Changeover times are manufacturing line dependent. Preventive maintenance times are considered which reduces the production available time and differentiate the available capacity from one period to another, Demand must be satisfied without backlogging. The goal of production planning is to determine optimal lot sizes and the production sequence of these on different manufacturing lines minimizing set-up and inventory costs over a planning horizon, Its not about finding a solution that purely satisfies demand on time, but finding a production plan integrating lot-sizing and scheduling with minimal cost. As shown above, this manufacturing process can be modeled by the lost-sizing and scheduling problem, so we studied and reformulated models of lot-sizing and scheduling models on literature. The capacitated simultaneous lot-sizing and scheduling on parallel non identical machines problem presented in this thesis is classified as a large bucket, multi-item, multi-machines, single-level, static and deterministic. The periods are differentiated in macro and micro periods. Macro periods are subdivided into many micro periods. Here, various types of printers can be produced in each macro period. The sequence decision appears on determining the best allocation of production lots in each micro period. The thesis consists on the development and implementation of a mathematical model of optimization to help on production planning managers decision process. The model integrates tactical decisions of lot sizing and operational decisions of sequencing these lots. It covers a planning horizon of 96 micro periods in 2 months and determines not only the size and the moment to produce a lot, but also the best sequence and allocation of these lots on manufacturing lines. The problem are considered NP-hard (see Bitran & Yanasse [7], so we develop a strategy of addition of families of inequalities and some reformulated constraints to allow this practical lot-sizing and scheduling problem to be effectively solved via mixed-integer programming and general MIP software incorporating cutting planes such as XPRESS and CPLEX. The models parameters such as demand, productivity of each line for each printer, available time per shift, preventive maintenance data and duration, inventory and set-up costs are deterministic and were collected with the process coordinators engineers. The data were analyzed and treated to be solved with MIP solver. The proposal resulting plan is than compared with the current practice, Improvements on productivity and costs reduction are evidenced. The new plan also increases the quality of planning, reducing to zero the number of delays in delivery of products. A recommended evaluation is the development of software with users graphical interface, which is strategic to the company because it allows a satisfactory use by production managers with little knowledge of optimization (Fleischmann et al.[15].