Aplicação de algoritmos genéticos para o problema de escalonamento de tarefas em sistemas de manufatura com controle supervisório e autômatos com parâmetros

Abstract

The job-shop scheduling aims to find optimal sequences of events that increases productivity indexes in a manufacturing systems. This work employs such metaheuristics to obtain optimal scheduling solutions, combined with the supervisory control theory (SCT) in the automatically generation of sequences of operations involved in the production of parts in a didactic manufacturing system. The SCT allows deriving control structures (supervisors) that formally guarantee the plant operation correctness and safeness. In turn, the genetic algorithm searches for the best possible sequence of events, among all enabled by the supervisors, exploring the plant parallelism, allowing to meet production goals while complying with safety and operational specifications. An example based on a didactic manufacturing cell illustrates the used approach where it is possible to observe the improvement of the task scheduling compared to sequential scheduling. Three different representations for the chromosome were used, which are the operation-based representation, random keys representation and priority rule-based representation. The obtained results in the case study simulation demonstrate the efficiency in the use of the job-shop-scheduling based in genetic algorithms with the use the SCT for description of the system restrictions.