Loss-detection model with Kanban card and its performance considering mixed loss

In modern assembly lines, it is commonly observed that a worker chooses some adequate part from several parts racks and assembles a product. To enhance performance and quality of such processes, it is important to confirm that an adequate part is selected. To address that problem, we proposed a loss-detection model using Kanban cards. However, the model presents some problem such as types of loss occurrences. As described herein, we present a new discrete-time Markov Chain model to describe the loss types. The model performance is presented using numerical experiments. Furthermore, results show that undetected-loss risk becomes the maximum when the transition probabilities are identical.