Proposta de solução de problemas de scheduling considerando possibilidade de terceirização usando a técnica de otimização por colônia de formigas

• Ano de defesa: 2010
• Autor(a) principal: Tavares Neto, Roberto Fernandes
• Orientador(a): Godinho Filho, Moacir
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Carlos
• Programa de Pós-Graduação: Programa de Pós-Graduação em Engenharia de Produção - PPGEP
• País: BR
• Palavras-chave em Português: Programação da produção; Sequenciamento da produção; Terceirização; Formiga - comportamento - modelos matemáticos; Inteligência artificial; Metaheurísticas; ACO
• Palavras-chave em Inglês: Scheduling; Ant colony otimization; Outsourcing
• Área do conhecimento CNPq: ENGENHARIAS::ENGENHARIA DE PRODUCAO::ENGENHARIA DO PRODUTO
• Link de acesso: https://repositorio.ufscar.br/handle/20.500.14289/3358

Althought the scheduling-related literature has a high level of diversity, just a small group have been considering the possibility of outsource a set of tasks. During a literature review, only two papers related to this theme were found, both dealing on scheduling projects with outsource possibilities on single-machine environments. Along with this scenario, it was possible to stablish the ACO (Ant Colony Optimization) algorithm as a promissing tecnique to solve combinatorial problems, including scheduling problems. This thesis approaches two scheduling problems with outsourcing allowed: (i) a scheduling problem in single machine manufacturing environment and (ii) a scheduling problem in a flowshop environment. For each problem, a new ACO algorithm is proposed and implemented. To verify the quality of the results, are also proposed and implemented: (i) a mathemetical programming model for the single machine environment problem; (ii) a branch and bound algorithm for the single machine environment problem and (iii) a a mathemetical programming model for the flowshop environment problem. The results shown that both ACO algorithms generate close-to-optimal results in a shorter computational time. In the case of the single machine environment problem, the presented results are best than the results related on the literature.