Evaluating the effect of learning rate, batch size and assignment strategies on the production performance

Task assignment methods usually rely on the fixed mean processing times of operations with the intent of balancing the workload assigned to operators or workstations in the production line. This assignment usually neglects the variability of operator processing times. In this work, a methodology in which the time in which an operator executes a task is variable, accordingly to a learning model, is proposed. It is exploited in order to assess the real-time task assignment adopted in the actual factory. The results show that, by including a learning model, it is possible to predict more accurately the long-term cycle time of the process. Standard scheduling strategies (first operator available, the operator closest to the machine) were compared with learning-oriented strategies (the most skilled, the least skilled). Through the case study, the paper addresses the problem of using a dynamic task assignment.an illustration.