A Gravitation-Based Hierarchical Community Detection Algorithm for Structuring Supply Chain Network

Abstract

AbstractAs industrial production outsourcing expands, the collaboration relationship of firms evolves to be more entangled, which means that the enterprise communities in the supply chain network become increasingly overlapping and their boundaries are ambiguous. Given the network complexity, deeper insight into the sequencing orders of suppliers and assemblers is required to orchestrate the supply chain partner collaboration. Considering the overlapping community and multi-layered connectivity characteristics of the supply chain network, in this paper, we design a gravitation-based hierarchical community detection algorithm for structuring the supply chain network. The solution applies a functional modules identification strategy based on node gravitation and a hierarchical clustering strategy based on module gravitation to structure the supply chain network architecture. The key technique is to investigate the global gravitational influence of focal firms, segment the functional modules by characterizing the overlapping conditions among communities, and construct the dendrogram by measuring the gravitational forces between modules in order to map the hierarchical architecture of the dendrogram to structure the supply chain network. The proposed algorithm does not necessitate a prior knowledge about the network. It is adaptable to construct the supply chain network that exhibits scale-free, highly overlapped modular community, and hierarchical characteristics. Experimental results on synthetic benchmark and real-world networks demonstrate the effectiveness and applicability of the proposed algorithm.