A Similarity-Based Hierarchical Clustering Method for Manufacturing Process Models

Abstract

As the adoption of information technologies increases in the manufacturing industry, manufacturing companies should efficiently manage their data and manufacturing processes in order to enhance their manufacturing competency. Because smart factories acquire processing data from connected machines, the business process management (BPM) approach can enrich the capability of manufacturing operations management. Manufacturing companies could benefit from the well-defined methodologies and process-centric engineering practices of this BPM approach for optimizing their manufacturing processes. Based on the approach, this paper proposes a similarity-based hierarchical clustering method for manufacturing processes. To this end, first we describe process modeling based on the BPM-compliant standard so that the manufacturing processes can be controlled by BPM systems. Second, we present similarity measures for manufacturing process models that serve as a criterion for the hierarchical clustering. Then, we formulate the hierarchical clustering problem and describe an agglomerative clustering algorithm using the measured similarities. Our contribution is considered on the assumption that a manufacturing company adopts the BPM approach and it operates various manufacturing processes. We expect that our method enables manufacturing companies to design and manage a vast amount of manufacturing processes at a coarser level, and it also can be applied to various process (re)engineering problems.