Metaheurísticas para o problema de Flowshop flexível com penalidades de adiantamento e atraso

• Autor(a) principal: Pereira, Ana Amélia de Souza
• Data de Publicação: 2011
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: LOCUS Repositório Institucional da UFV
• Texto Completo: http://locus.ufv.br/handle/123456789/2623
• topic: Programação de tarefas; Metaheurísticas; Otimização combinatória; Job scheduling; Metaheuristics; Combinatorial optimization; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO

This work deals with the problem of job scheduling in flexible flowshop system with the objective of minimizing the penalties for tardiness and earliness in terms of the job due date. It is considered that all jobs are available for processing in different moments, according to them release times, besides that, there are setup times depending on the sequence and stages. The problem addressed belongs to the production programming environment JIT (just-in-time). In this environment, each job has a processing time and due date within which it should preferably be completed, therefore the processing a job should end as closer as possible to its due date. When a job finishes processing earlyth, production is finalized before due date, hence it must be stored, which generates storage costs. When the job is tardy is the job and consequent delay on further productions, causing a chain effect, causing loss of trust from consumer, which has unestimated cost, besides contract fines, which harms the image of an enterprise. In this work it was developed an PLIM for the problem. Due to the problem computational complexity, heuristics and metaheuristic are used for obtaining good quality approximate solutions. Metaheuristics applied are Iterated Local Search (ILS) and Genetic Algorithm (GA): basic genetic algorithm (GA-B) and Genetic Algorithm with local search (GA-LS). The GA-LS is an adaptation of the GA-B with local search, applied for improving a solutions determined by the genetic operators. PLIM model, for of problems smaller size, is solved by using the optimization software CPLEX. Results obtained from the proposed heuristics ILS, AG-B and AG-BL are compared and analyzed among themselves and also with solutions determined by constructive heuristics, MST and NEH-MST, computational tests have shown that metaheuristic ILS obtains solutions of good quality. After adjusting the required parameters and several tests showed that the ILS was the most effective technique for obtaining good quality solutions to the problem studied.