Structural Complexity Of Industrial Ecosystems From The Complex Network Perspective

Many misconceptions and mistakes in the eco-industrial practices can be chalked up to lack of knowledge of complexity in industrial ecosystems. The complexity embodied in industrial ecosystems comprises of component complexity, structural complexity and functional complexity which are coupled and interdepend with each other. The uncovering of structural complexity would be helpful for us to understand how industrial ecosystems evolve and what interaction patterns exist between their structures and functions. Based on complexity sciences and complex networks, an analytical framework was proposed to uncover the organizational and hierarchical network characteristics in industrial ecosystems. And the corresponding physical meanings and policy implications were explored. Four main contributions were made in this dissertation.Firstly, a framework of structural complexity for industrial ecosystem was built. Based on the understanding of the essence of industrial symbiosis and industrial ecosystem, three kinds of industrial networks were abstracted: production network, physical exchange network(by-products and product relations included), and industrial symbiosis network(utility sharing and physical exchange relations included). Inspired by the classification of ontological complexity suggested by Rescher, an analytical procedures of structural complexity based on network theory was proposed.Secondly, case studies of industrial ecosystems were collected and their component complexity was analyzed. Component complexity was the foundation of structural complexity, composed by constitutional and taxonomical complexity. We collected 38 case studies published in SCI and CNKI databases and finally choosed 13 cases with the guide of proposed prinicples of data filtering. 15×3 industrial networks in total were reconstructed from the common cognition of industrial ecosystems. Obvious differences were observed in the component complexity between product networks and each of the other two network types. Many bidirectional exchange relations between production units emerged with the addition of by-product exchange behaviors and also utility sharing relations.Thirdly, motif identification method was built for small-scale networks and organizational complexity of industrial ecosystems was analyzed. Critical upper and lower boundaries of network sizes suitable to analyze subgraphs and motifs were identified based on Erdos-Renyi random network model. For production networks, physical exchange networks and industrial symbiosis network, the number of high frequency triads were 3, 6 and 7 respectively, and 8, 15 and 18 for high frequency tetrads. Both of these findings meant that the organizational forms of physical exchange networks and industrial symbiosis networks were much more diversified than that of product networks.Fourthly, the analysis of hierarchical complexity suggested that product networks showed obvious hierarchical feature, while the inclusion of physical exchanges and utility sharing relations tended to connect upstream and downstream production units and formed more complex structures. Meanwhile, these two types of symbiotic relations formed some close-loop circulations of materials within production systems, which diversified the chains of material and energy transfer in industrial symbiosis networks.In conclusion, the analytical method of structural complexity based on complex network theory contributes the transition of industrial ecology from ecological metaphor to an explicit, measurable and comparable scientific paradigm, which is also the first step towards a theory of structure-performance relationship of industrial ecosystem and a more solid industrial ecology theory.