CLUSTERING ORGANIZATION STRUCTURE IN PRODUCT DEVELOPMENT PROJECTS USING SIMILARITY

A product development (PD) project is a complex social network, in which teams have their own attributes and are related by information flow. Similar team attributes and the complex patterns of technical dependency among teams both affect organization modularity. This paper provides an innovative spectral clustering approach that merge team attributes and relationship of teams. To measure the similarity of PD teams, we analyze the similar attributes of team and build structural models to capture the technical communication dependency among teams via the product-organization multi-domain matrix (MDM). We use two metrics to evaluate the clustering solutions and confirm that the proposed approach provides effective reduction of PD coordination complexity.