Abordagem para otimização multiobjetivo de regras heurísticas de sequenciamento em sistemas de manufatura job shop por meio de simulação computacional acoplada ao algoritmo genético

Abstract

Since the arising of the Toyota Production System (TPS), known as lean production, and more recently the consumer market demands, manufacturing companies have worked to improve their production processes in order to reduce inventory levels, to eliminate waste and to maintain quality, competitiveness and profit. In this context, the sequencing of production orders is one of the most complex problems facing businesses and it is object of several studies. Thus, this paper presents an alternative approach in resolving this problem, i.e.: to use a simulation model as the objective function in genetic algorithm instead analytical mathematical equation. So, for each situation, will not need to change the equation, but adjust the model and to make a new simulation. Heuristics sequencing rules in job shop environments were considered, with routes, due dates and times of operation generated randomly, in order to determine the best programming technique for performance in relation to the total time of crossing, the total tardiness and the number of tardy jobs. Results corroborate the method adopted. Multi-objective optimization approach is based on the coupling of Genetic Algorithm with an Arena simulation model through the Visual Basic for Application language and the ActiveX Automation controller.