An empirical investigation of learning curve composition laws for quality improvement in complex manufacturing plants

Learning behaviors related to quality improvement in manufacturing systems (i.e. reduction of defectiveness over production cycles) are widely investigated. Many different approaches have been introduced to describe the link between the learning mechanism and quality performance of a plant. In a previous study by the same authors, a set of learning “composition laws” for two basic structures were defined to provide a tool to forecast the behavior of complex manufacturing systems composed by a network of elementary processes. This paper presents an empirical investigation about these learning composition laws on a real case in the field of automotive exhaust‐systems manufacturing.