Analytical Modeling of Human Choice Complexity in a Mixed Model Assembly Line Using Machine Learning-Based Human in the Loop Simulation

Despite the recent advances in manufacturing automation, the role of human involvement in manufacturing systems is still regarded as a key factor in maintaining higher adaptability and flexibility. In general, however, modeling of human operators in manufacturing system design still considers human as a physical resource represented in statistical terms. In this paper, we propose a human in the loop (HIL) approach to investigate the operator's choice complexity in a mixed model assembly line. The HIL simulation allows humans to become a core component of the simulation, therefore influencing the outcome in a way that is often impossible to reproduce via traditional simulation methods. At the initial stage, we identify the significant features affecting the choice complexity. The selected features are in turn used to build a regression model, in which human reaction time with regard to different degree of choice complexity serves as a response variable used to train and test the model. The proposed method, along with an illustrative case study, not only serves as a tool to quantitatively assess and predict the impact of choice complexity on operator's effectiveness, but also provides an insight into how complexity can be mitigated without affecting the overall manufacturing throughput.