Algoritmo baseado em colônia de abelhas artificiais para resolução do problema de programação de um job shop flexível multiobjetivo

• Ano de defesa: 2019
• Autor(a) principal: Florêncio, Guilherme Felipe
• Orientador(a): Kato, Edilson Reis Rodrigues
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Carlos; Câmpus São Carlos
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ciência da Computação - PPGCC
• Palavras-chave em Português: Programação da produção; Multiobjetivo; Colônia de abelhas artificiais; Pareto
• Palavras-chave em Inglês: Production program; Multiobjective; Artificial bee colony; Flexible job shop problem (FJSP)
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::METODOLOGIA E TECNICAS DA COMPUTACAO; CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::SISTEMAS DE COMPUTACAO
• Link de acesso: https://repositorio.ufscar.br/handle/20.500.14289/11793

The flexible job shop problem (FJSP), considered one of the most complex problems of work scheduling, composes the class of NP-hard problems in the computer science field. Although this scheduling production problem is quite complex, it can be computationally more costly to be treated when additional restrictions or criteria are imposed to the problem. The FJSP consists in a set of "n" job that are constituted by "i" operations, and these operations are processed individually by a machine 'M' that is part of the production environment machine set. In this type of system, each operation inside the job can be processed at a different machine. On the fully flexible process, all the machines are able to process all the operations, and on the partly flexible, at least one does not process at least one operation. This problem can be divided into two sub-problems, the routing and the scheduling. The routing consists in to define which machine will process the operation, and the scheduling consists in the order that operations will be processed. Through this work, it has been aimed to minimize performance multiobjectives, such as: makespan, more loaded machine load and full load of all machines, seeking high diversity of solutions. To achieve the objectives of this paper it was implemented the metaheuristic Artificial Bee Colony along with the Pareto method to help treating the multiobjectives. The observed results were satisfactory in most instances for which the algorithm was applied, and optimal results were found for some instances.