Este trabalho aborda o problema de programação da produção em um flexible flow line com tempos de setup. De acordo com a literatura, este ambiente pode ser considerado como um caso especial do Flow Shop com múltiplas máquinas, onde as tarefas podem saltar estágios. Neste estudo, foram analisados dois problemas: o primeiro, com tempos de setup independentes da sequência, e o segundo, com setup dependente da sequência de tarefas. Além disso, o setup das máquinas para as tarefas pode ser antecipado ou não. No primeiro caso, as máquinas de um estágio podem ser preparadas para o processamento de uma tarefa antes do seu término no estágio anterior. Se o setup não pode ser antecipado, a tarefa deve esperar o seu término no estágio de produção anterior. Este ambiente produtivo pode ser encontrado em um vasto número de indústrias tais como química, eletrônica, automotiva e têxtil. A medida de desempenho dos problemas é a duração total da programação (makespan). Este é um critério apropriado para sistemas de produção com grandes cargas de trabalho e em que a utilização dos recursos produtivos em longo prazo deve ser otimizada. O exame da literatura mostrou que há poucos estudos abordando a programação em flexible flow line. Considerando este aspecto, este trabalho apresenta heurísticas construtivas originais para a obtenção de programações apropriadas ao problema mencionado. Uma extensiva experimentação computacional foi executada para avaliar o desempenho relativo das heurísticas. Os resultados experimentais foram analisados e discutidos. / This work addresses the job scheduling on a flexible flow line with separate setup times. According to the literature, this scheduling problem can be considered as a special case of the Flow Shop with multiple machines, where the jobs may skip stages. Two modeled problems have been studied. In the first scheduling problem the setup times are sequence independent, and in the second one these times are sequence dependent. Moreover, the machine setup task can be either anticipatory or non-anticipatory. In the first case, a k-stage machine may be prepared for a job processing before its completion on the k-1 production stage. Otherwise, the setup task must wait for the job completion on the former production stage. This production environment can be found in a number of industries such as chemicals, electronics, automotive, and textiles. The performance measure of the production schedules is the makespan, that is, the total time to complete the schedule. This is an appropriate performance criterion for production systems with large workloads, and where the utilization of productive resources in the long term should be optimized. The literature examination has shown that there is a small number of studies dealing with flexible flow line scheduling. Having this in mind, this work introduces original constructive heuristics in order to obtain suitable schedules for the aforementioned scheduling problem. An extensive computational experience has been carried out in order to evaluate the relative performance of the heuristics. Experimental results are discussed.

Este trabalho aborda o problema de programação da produção em um flexible flow line com tempos de setup. De acordo com a literatura, este ambiente pode ser considerado como um caso especial do Flow Shop com múltiplas máquinas, onde as tarefas podem saltar estágios. Neste estudo, foram analisados dois problemas: o primeiro, com tempos de setup independentes da sequência, e o segundo, com setup dependente da sequência de tarefas. Além disso, o setup das máquinas para as tarefas pode ser antecipado ou não. No primeiro caso, as máquinas de um estágio podem ser preparadas para o processamento de uma tarefa antes do seu término no estágio anterior. Se o setup não pode ser antecipado, a tarefa deve esperar o seu término no estágio de produção anterior. Este ambiente produtivo pode ser encontrado em um vasto número de indústrias tais como química, eletrônica, automotiva e têxtil. A medida de desempenho dos problemas é a duração total da programação (makespan). Este é um critério apropriado para sistemas de produção com grandes cargas de trabalho e em que a utilização dos recursos produtivos em longo prazo deve ser otimizada. O exame da literatura mostrou que há poucos estudos abordando a programação em flexible flow line. Considerando este aspecto, este trabalho apresenta heurísticas construtivas originais para a obtenção de programações apropriadas ao problema mencionado. Uma extensiva experimentação computacional foi executada para avaliar o desempenho relativo das heurísticas. Os resultados experimentais foram analisados e discutidos. / This work addresses the job scheduling on a flexible flow line with separate setup times. According to the literature, this scheduling problem can be considered as a special case of the Flow Shop with multiple machines, where the jobs may skip stages. Two modeled problems have been studied. In the first scheduling problem the setup times are sequence independent, and in the second one these times are sequence dependent. Moreover, the machine setup task can be either anticipatory or non-anticipatory. In the first case, a k-stage machine may be prepared for a job processing before its completion on the k-1 production stage. Otherwise, the setup task must wait for the job completion on the former production stage. This production environment can be found in a number of industries such as chemicals, electronics, automotive, and textiles. The performance measure of the production schedules is the makespan, that is, the total time to complete the schedule. This is an appropriate performance criterion for production systems with large workloads, and where the utilization of productive resources in the long term should be optimized. The literature examination has shown that there is a small number of studies dealing with flexible flow line scheduling. Having this in mind, this work introduces original constructive heuristics in order to obtain suitable schedules for the aforementioned scheduling problem. An extensive computational experience has been carried out in order to evaluate the relative performance of the heuristics. Experimental results are discussed.

Submitted by Liliane Ferreira (ljuvencia30@gmail.com) on 2018-07-09T15:51:41Z No. of bitstreams: 2 Dissertação - Jeferson Silva Martins - 2018.pdf: 4962348 bytes, checksum: 7da8c895b4929a650a07c2381fc59290 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2018-07-10T11:07:38Z (GMT) No. of bitstreams: 2 Dissertação - Jeferson Silva Martins - 2018.pdf: 4962348 bytes, checksum: 7da8c895b4929a650a07c2381fc59290 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2018-07-10T11:07:38Z (GMT). No. of bitstreams: 2 Dissertação - Jeferson Silva Martins - 2018.pdf: 4962348 bytes, checksum: 7da8c895b4929a650a07c2381fc59290 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2018-05-17 / This paper presents the Mixed-Model Flow Line problem for Production Scheduling. Production problems have been vastly studied given their usability in all kind of assembly lines and its production in the industry worldwide. The problem processes jobs grouped by families in stations of machines seeking for the best sequencing the families can provide to be produced in the processing time (Makespan). In this work, a bibliographic review was performed to support the study, looking different methods and problems that encompass the Mixed-Model Flow Line in the literature showing its importance towards the industrial and scientific environment. The Mixed-Model Flow Line problem was solved through optimization techniques: An integer lienar programming inGUROBI, Lagrangian Relaxation and Genetic Algorithm. The processing times and demands are randomly generated for each instance. Finally, the results obtained in the techniques used for the problem were compared to each other. / Este trabalho apresenta o problema de Flow Line Misto para a Programação da Produção em Scheduling. Problemas da produção tem sido vastamente estudado dada sua usabilidade em todo o tipo de linha de montagem e produção na indústria mundial. O problema em estudo considera o processamento de tarefas em conjuntos por meio de estações de máquinas buscando um bom sequenciamento das tarefas a serem produzidas melhorando o tempo de processamento (Makespan). Neste trabalho, uma revisão bibliográfica foi realizada para fundamentar o estudo encontrando diferentes métodos e problemas que englobam o Flow Line Misto mostrando a importância deste estudo no meio industrial e científico. O problema do Flow Line Misto foi resolvido por meio de técnicas de otimização, sendo elas um modelo de programação linear inteira resolvido pelo GUROBI, no Lagrangeano Relaxado e no Algoritmo Genético com tempos de processamento e demandas gerados aleatoriamente para as instâncias criadas. Por fim, os resultados obtidos nas técnicas analisadas para o problema foram comparados entre si.

This paper presents a model and analysis of a synchronous tandem flow line that produces different part types on unreliable machines. The machines operate according to a static priority rule, operating on the highest priority part whenever possible, and operating on lower priority parts only when unable to produce those with higher priorities. We develop a new decomposition method to analyze the behavior of the manufacturing system by decomposing the long production line into small analytically tractable components. As a first step in modeling a production line with more than one part type, we restrict ourselves to the case where there are two part types. Detailed modeling and derivations are presented with a small two-part-type production line that consists of two processing machines and two demand machines. Then, a generalized longer flow line is analyzed. Furthermore, estimates for performance measures, such as average buffer levels and production rates, are presented and compared to extensive discrete event simulation. The quantitative behavior of the two-part type processing line under different demand scenarios is also provided. / Singapore-MIT Alliance (SMA)

Le but de ce travail était la modélisation mécanique et thermique du laminage asymétrique. Celle-ci a été effectuée à l'aide de diverses méthodes. Deux approches, l'une analytique et l'autre semi-analytique, ont été utilisées pour la modélisation mécanique et deux autres, celles des différences finies et des éléments finis, pour la modélisation thermique. La première approche est celle des champs uniformes ; on a développé un modèle simple qui peut donner des ordres de grandeurs des différents paramètres du procédé. Une deuxième méthode plus complète mais plus complexe se base sur un champ de vitesse déduit d'une estimation parabolique de la forme des trajectoires. L’avantage de ce modèle est qu’il est capable de prévoir des grandeurs locales. Enfin la méthode des éléments finis a été aussi utilisée pour la modélisation mécanique. Dans un deuxième temps, l’aspect thermique du laminage asymétrique a été étudié. Pour la modélisation du procédé, nous avons alors adopté deux méthodes numériques, les différences finies et les éléments finis, du fait de la complexité de celui-ci. Le modèle thermique développé par différences finies a été ensuite intégré dans notre modèle mécanique (méthode des trajectoires). Celui-ci a donc pu être étendu pour pouvoir utiliser des lois d’écoulement qui dépendent de la température et donc pouvoir simuler le laminage asymétrique à chaud. Les prédictions du champ de température ont été confrontées aux résultats de calculs par éléments finis obtenus avec le code Abaqus. Finalement, une étude simple du développement de la texture cristallographique au cours du laminage a été réalisée. / The aim of this work was to propose a mechanical and thermal modelling of asymmetric rolling. This was achieved using various methods. Two approaches, the first one analytical and the second semi-analytical, were employed for mechanical modeling. Two others methods, the finite difference and finite element methods, were used for thermal modeling. The first analytical model is an upper bound method based on a uniform strain field; a simple model was developed that can give orders of magnitude of the various process parameters. The second model uses a more refined analytical velocity field based on the classical parabolic estimation of the material flow lines in rolling. The advantage of this model is its ability to provide local quantities. Finally, the finite element method was also used for mechanical modeling.The thermal aspects of asymmetric rolling were investigated in turn by two methods. For modeling the process, we adopted two numerical approaches, the finite difference and finite element methods, because of the complexity of the latter. The thermal model developed by finite differences was integrated into our mechanical model (flow line model). It was then extended to investigate asymmetrical hot rolling. The predictions of the temperature field were compared with results from finite element calculations obtained with the Abaqus code. Finally, a simple study of crystallographic texture development was achieved.