Um modelo de otimização baseado em algoritmo memético para o escalonamento de ordens de produção utilizando divisão de lotes de tamanho variável

Abstract

The contribution of meta-heuristics, especially evolutionary algorithms, in combinatorial optimization area is extremely important, as they help in finding near optimal solutions to complex real-life problems whose resolution is infeasible in acceptable time due to its computational complexity, offering an important flexibility in the modeling of problem. This study propose to present and implement a computational model to be used in optimizing the production scheduling of manufacturing orders using a Memetic Algorithm that allows to search both the best sequence of jobs as of variable size batches that the quantity of each operation can be subdivided. The possibility of using alternative resources, operations with secondary resources, unavailability intervals and batch transfer lots are features presented in the model, which lends it great robustness and applicability to flexible manufacturing environments, allowing the modeling of Flexible Job Shop Scheduling Problem (FJSSP) that reflects with higher accuracy the real manufacturing environment, generating optimized scheduling results that are adhering to the plant needs. Multiple instances of FJSSP are used in the tests and the results show that the proposed algorithm succeeds in optimizing the scheduling of production orders for each instance so efficient.