Theoretical And Empirical Study On Regional Industrial Structure Based On Complex Network Theory

The21st century has been called the complexity of the century by the famousphysicist Stephen Hawking. The complex network research is an important part of thecomplexity research. Over the past decade, complex network is considered as the highhopes as a powerful tool to study the complexity science and complex systems. Thereality of complex social networks showing some common topology and properties ismore and more concerned. And the results achieved on complex social networks arepromoting the development of complex network theory faster and faster. The complexsocial networks are those social networks with complex network characteristics. Thetheoretical framework of complex social network is composed of microscopic mainbody dynamics behavior and the main interaction between the common emerged theoverall dynamic behavior, so using the complex network theory to study the regionalindustrial structure is very suitable.Industrial economic system is a typical complex system. Based on the complexnetworks theory, if the industrial departments are the nodes in the network and therelationship of industries’ inputs and outputs of are the edges between the nodes, thewhole industry economic system is a complex network. On the basis of relatedresearches, this paper established the regional industrial correlation network modelbased on the complex networks theory. This research provides a new researchperspective for regional industrial structure. In this paper, the completed work andmain innovative points as follows:(1)Introduced the complex network theory to compare and analysis regionalindustrial structure, and the method of complex networks is improved to be moresuitable to the weighted, directed and having self-loops networks.(2) Enriched and developed the complex network theory. On the basis of theresearch of the domestic and foreign scholars we introduced the concept of thresholdand discussed the network’s topology in the case of setting different thresholds. In thisway, we eliminated the influence of weak relationships between industries to makesthe results more clearly. Therefore the results have a stronger guiding significance.(3)An improved module function is proposed to analysis the weighted networks with having characteristics of overlapping community. The improved module functionis used to the empirical research. The results provide necessary theoretical support forthe government to adjust industrial structure and make full use of the regionaladvantages. Demographic characteristics of the industry association networks ofBeijing, Shanghai and China are analysed. The paper also provides a new algorithmfor partitioning complex network into communities. Using the new algorithm theindustry association networks of Beijing, Shanghai and China are partitioned topropose practical policy and recommendations.