Interdisciplinary assessment of product similarity by measures for network similarity

01.07.2019 bis 30.06.2020

Evaluating the performance of a production system is an important aspect of production improvements, which is based on key performance indicators (KPIs). As many of the KPIs are product dependant, the more product types are produced, the more KPIs need to be regarded. The objective of the project ProNetSim is to reduce the complexity of evaluating real-world production systems by clustering similar products, which leads to reduction in the number of KPIs to be evaluated. Uncovering the similarity of products is not an easy task. Our principle strategy is to represent the relevant criteria and KPIs, including the flexibilities in the sequence of process steps, in terms of a directed network. We propose to model each product as a directed network, whose nodes represent the machines in a production line and possess attributes denoting the KPIs of the machine for the particular product type. An exhaustive study is necessary to evaluate the applicability of different network similarity or distance measures in the context of production systems.
Thus, the analysis of network similarity can highly contribute to the performance evaluation of a production system.