masterThesis
Método baseado em inteligência artificial para previsão do prazo de entrega de tarefas em estações de manufatura
Fecha
2020-10-30Registro en:
MODESTI, Paulo Henrique de. Método baseado em inteligência artificial para previsão do prazo de entrega de tarefas em estações de manufatura. 2020. Dissertação (Mestrado em Engenharia Mecânica e de Materiais) - Universidade Tecnológica Federal do Paraná, Curitiba, 2020.
Autor
Modesti, Paulo Henrique de
Resumen
An important competitive advantage of an industry is the ability to meet the delivery times promised to its customers. However, with the increase in demand for customized products, imposed by the market in the last decades, the complexity to estimate manufacturing times increases. This difficulty is amplified in some sectors of the industry where a great variety of parts are produced and whose demand for new tasks is random. Examples of this sector are companies such as job shops, prototype manufacturers and spare parts, as they operate with low volumes of production per part and have to deal with variable lead times and the need for quick presentation of delivery times. Thus, it is evident that the predictability of the delivery time is a complex process. In this way, the objective of this research was defined as the development of a method, capable of predicting the deadline for the delivery of tasks in real time, aiming to assist in decision making regarding the planning of industries. For the development of this method, Design Science Research was used as the methodological framework. Thus, six steps were taken: (i) problem identification and motivation; (ii) definition of the solution’s objectives; (iii) design and development; (iv) demonstration; (v) evaluation and (vi) communication of results. To demonstrate the method, the same was applied in a real case, in the prototypes department of an appliance company in the region of Curitiba, being evaluated through questionnaires and comparisons. When analyzing the results, it is proved that the proposed method allows greater assertiveness in the task delivery date forecasts. This work focused on the study of a prototype department of a company in the white good business, but there is a possibility that the model could be adapted to other segments.