Bottleneck prediction using the active period method in combination with buffer inventories

Abstract

Knowing the bottleneck is one of the keys to improving a production system. The active period method is one approach to detect shifting bottlenecks that most other bottleneck detection methods have problems with. Yet, like many other methods, these detections are limited to detecting the past and present bottlenecks. In this paper, we combined the active period method with the buffer inventories and free buffer spaces of the adjacent inventories to statistically predict not only an upcoming change of the bottleneck, but also where the bottleneck will move to.