Contributions to new variants of the open shop scheduling problem: modeling and solution methods

• Autor(a) principal: Abreu, Levi Ribeiro de
• Data de Publicação: 2023
• Tipo de documento: Tese
• Idioma: eng
• Título da fonte: Biblioteca Digital de Teses e Dissertações da USP
• Texto Completo: https://www.teses.usp.br/teses/disponiveis/18/18156/tde-19052023-162053/
• topic: open shop; algoritmos de aproximação; algoritmos populacionais; approximation algorithms; mathematical modeling; modelagem matemática; open shop; population algorithms

Several studies have been carried out regarding optimizing production scheduling in industrial environments. As a result, new variants related to several problems have been incorporated into this study area to cover the most diverse cases presented in productive environments. In this sense, this dissertation aimed to study a production scheduling problem little stressed in the literature, the production scheduling problem in an open shop environment with many applications in the industrial and services areas. In this study, we observed essential constraints for the described environment, which became new variants for the problem: the sequence-dependent setup times; the study of blocking machines in the processing of operations; the study of reprocessing or repetition of operations in the production process and the delivery of products through the vehicle routing. The objective of these problems was to minimize the total duration of the schedule (makespan). We proposed heuristic methods for modeling and solving these problems, such as priority rules, constructive techniques, bio-inspired meta-heuristics, and mathematical programming methods such as integer linear programming and constraint programming models and matheuristics. For the computational tests, we ran the methods with robust data from classical literature instances adapted to the constraints of the problems under consideration and new instances proposed during the study. The results showed that the proposed exact and approximate methods provided quality solutions with computational efficiency and were competitive compared to the literature methods.