Combining workers' behavior data and real time simulator for a cellular manufacturing system

This paper presents a real time simulator using workers' behavior data to support a product manager's decision making in cellular manufacturing system. The real time simulator is a simulation using the data collected in real time, which is effective in reduce prediction error and support quick decision making. However, due to a high analytical cost in the conventional work measurement, it is difficult to continuously collect the worker's behavior data of the cellular manufacturing system in real time. For this purpose, first we built an experimental cellular manufacturing line that is able to collect the workers' behavior data in real time using the Bluetooth low energy beacons and acceleration sensors. Thereby, we constructed a real time simulation to combine the analyzed behavior data and a discrete event simulation. The result revealed that, the relative error between our method and the video observation method is small, and the prediction error of our simulation is smaller than using the standard work time.