Método de escolha de Scheduling para problemas de Job Shop Flexível utilizando um tomador de decisão Fuzzy

Abstract

Among the scheduling problems encountered in production, there is the flexible job shop (FJSP), which is an extension of the classic Job Shop. The FJSP is classified as an NP-Hard problem and can be described as a set of jobs, formed by a certain number of operations that can be allocated on a predetermined set of machines, with their respective execution times. The way operations are distributed across machines directly influences programming efficiency. Efficiency that can be measured and optimized through several performance criteria, that is, several objectives at the same time. The production schedule generated by a multi-objective optimization algorithm (MOFJSP), which mainly considers three performance criteria, namely: completion time of all operations (Makespan), load assigned to the most loaded machine and the sum of the load of all machines, does not present a single global optimal solution, but a set of non-dominated and dominated solutions, called the Pareto set. The solutions of this Pareto set are optimal or close to optimal solutions, being considered good solutions because they can generate a diversity of representations of the production schedule, for example, in the form of Gantt charts. For decision makers to choose the best production scheduling solution among the possible ones found by MOFJSP, other variables can be taken into account, such as maximizing or minimizing machine idleness, the load of operations on a machine, etc. which can provide greater adherence to the decision in view of the needs of the production system. For the inclusion of these variables and the selection of the best production schedule among those provided by the Pareto set, it is proposed to use a decision-making algorithm based on the Technique for Order of Preference by Similarity with the Ideal Solution in a Fuzzy environment, called Fuzzy - TOPSIS This decision maker can weight variables that are not contemplated in the MOFJSP algorithm and assist in the decision making of the best production schedule among the optimal ones or close to the optimal ones obtained. In the results, it was possible to notice that the values obtained with the application of the algorithm proposed for this problem obtained results close to the expected ones, according to the variables analyzed, and can be an important tool in aiding decision-making in a production system.