Loading...
Publication
Desenvolvimento de um sistema de apoio à decisão para planeamento e controlo de sistemas de produção do tipo Job Shop
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 4.84 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
A presente dissertação teve como objetivo desenvolver um sistema de apoio à decisão para o planeamento e controlo de sistemas de produção do tipo Job Shop, aplicado a uma PME do setor da metalomecânica. Este tipo de sistema produtivo, caracterizado por elevada flexibilidade e diversidade de produtos, apresenta desafios no cumprimento de prazos, na alocação de recursos e na redução de tempos de espera. A investigação iniciou-se com uma revisão bibliográfica sobre o modelo Job Shop, estratégias de planeamento e controlo, bem como o potencial das tecnologias da Indústria 4.0, heurísticas e algoritmos de inteligência artificial. Seguidamente, recorreu-se à metodologia Action Research (AR), que combinou a análise teórica com aplicação prática na empresa. Neste trabalho de investigação foi desenvolvida uma ferramenta de simulação para testar diferentes
regras de sequenciamento, destacando a heurística EDD+LPT (Earliest Due Date + Longest Processing Time), que visa conciliar o cumprimento de prazos de entrega com a redução de bloqueios provocados por operações longas. Os resultados demonstraram que a heurística EDD+LPT melhorou significativamente o planeamento da produção, reduzindo o atraso total e aproximando as datas previstas das datas pretendidas de entrega. Contudo, a variabilidade dos tempos de processamento e a dependência de operações externas continuam a gerar desafios, refletindo a complexidade do modelo Job Shop. Do ponto de vista prático, esta dissertação disponibiliza à empresa uma ferramenta prática de apoio à decisão, que permite aumentar a previsibilidade das entregas e a utilização dos recursos. Do ponto de vista acadêmico, evidencia a relevância da integração de heurísticas
simples com simulação, abrindo caminho para abordagens mais avançadas de otimização e inteligência artificial. Futuramente, sugere-se a integração de modelos probabilísticos, algoritmos avançados e monitorização em tempo real.
This dissertation aimed to develop a decision support system for planning and controlling Job Shop production systems, as applied to a small-to-medium enterprise (SME) in the metalworking sector. Such systems are characterised by high flexibility and product diversity, which makes it challenging to meet deadlines, allocate resources and reduce waiting times. The research began with a review of the literature on the job shop model, planning and control strategies, and the potential of Industry 4.0 technologies, heuristics and artificial intelligence algorithms. Next, Action Research was employed, combining theoretical analysis with practical application within the company. As part of this study, a simulation tool was developed to test different sequencing rules. This highlighted the EDD+LPT (earliest due date + longest processing time) heuristic, which aims to reconcile meeting delivery deadlines with reducing bottlenecks caused by long operations. The results showed that the EDD+LPT heuristic significantly improved production planning, reducing total delays and bringing the expected delivery dates closer to those desired. However, variability in processing times and dependence on external operations continue to pose challenges, reflecting the complexity of the Job Shop model. From a practical standpoint, this dissertation offers companies a valuable decision-making tool, increasing the predictability of deliveries and resource utilisation. From an academic perspective, it emphasises the importance of combining simple heuristics with simulation, paving the way for more sophisticated optimisation and artificial intelligence approaches. We suggest integrating probabilistic models, advanced algorithms and real-time monitoring in the future.
This dissertation aimed to develop a decision support system for planning and controlling Job Shop production systems, as applied to a small-to-medium enterprise (SME) in the metalworking sector. Such systems are characterised by high flexibility and product diversity, which makes it challenging to meet deadlines, allocate resources and reduce waiting times. The research began with a review of the literature on the job shop model, planning and control strategies, and the potential of Industry 4.0 technologies, heuristics and artificial intelligence algorithms. Next, Action Research was employed, combining theoretical analysis with practical application within the company. As part of this study, a simulation tool was developed to test different sequencing rules. This highlighted the EDD+LPT (earliest due date + longest processing time) heuristic, which aims to reconcile meeting delivery deadlines with reducing bottlenecks caused by long operations. The results showed that the EDD+LPT heuristic significantly improved production planning, reducing total delays and bringing the expected delivery dates closer to those desired. However, variability in processing times and dependence on external operations continue to pose challenges, reflecting the complexity of the Job Shop model. From a practical standpoint, this dissertation offers companies a valuable decision-making tool, increasing the predictability of deliveries and resource utilisation. From an academic perspective, it emphasises the importance of combining simple heuristics with simulation, paving the way for more sophisticated optimisation and artificial intelligence approaches. We suggest integrating probabilistic models, advanced algorithms and real-time monitoring in the future.
Description
Keywords
Job shop production planning EDD+LPT heuristics simulation artificial intelligence Planeamento da produção Heurísticas Simulação Indústria 4.0 Industry 4.0 Inteligência Artificial