At present, China’s economic growth is slowing from high-speed growth to medium-to-high speed of growth, and steady growth, structural adjustment, transfer mode and risk prevention are the development goals. Among them, how to adjust the industrial structure and promote the sustained and healthy development of the industry has become a key economic problem faced by our country. The circulation of material, energy and information among many industries has formed a complex industrial network. The network structure determines the function of the economic system and the direction of transformation and upgrading of the industrial structure. Therefore, in the new round of economic development, the study of China’s regional industrial network structure optimization problem provides a clear direction and task for the formation of efficient, coordinated and sustainable industrial rational distribution，and it has a very important theoretical and practical significance.
This paper uses the complex network theory, geographic information system, system dynamics and applied statistics to study optimization of regional industrial network structure based on link prediction. The research area is China’s industrial ecosystem. Because the regional industry network structure is very complex, in order to sort out the relationship, we should follow the research principle from simple to complex. This paper first compares and analyzes the regional industrial network structure heterogeneity between the whole (China) and the part (main function area) from the perspective of transverse static view. It paves the way for further research on the evolution path of regional industry network structure based on link prediction. Second, this paper studies the change characteristics of regional industrial network structure from vertical dynamic angle of view, and analyzes the optimization path of network structure from exploring the evolution course of the regional industrial network. Third, simulation research is carried out in this paper based on external impact timing link prediction model. The regional network structure optimization index is used to evaluate the simulation results. Then we get the optimized regional industry network structure and push the network structure optimization path. Finally, the internet and traditional industries are taken as examples and the paper studies the causal relationship between variables in link, and simulates the policy matching relation, then putting forward optimization strategy.
The main contents are as follows:
Firstly, this paper analyzes the characteristics of the regional industrial network structure. Taking into account the differences in the carrying capacity of regional resources in each subsystem, the paper focuses on the comparative study of the industrial eco economic system and the main functional areas. The chemical industry, the service sector including the postal, wholesale, retail and financial sectors, is very important in these two networks. The two networks differ in three variables with closeness centrality, betweenness centrality, and structural holes. The industry density of the whole country is higher than that of the main functional area, but the relationship between the two inputs and outputs is not strong. The industry of optimizing development district of “Beijing-Tianjin-Hebei” has a more important center position and industrial autonomy than the industry of key-developing district of the “Harbin long city group”. There is a single bridge link between the main functional areas, and the heavy industry of the optimizing development district has a supporting role to equipment industry of key-developing district. There is a high-energy industry included construction and transportation industry which compose “strong supply” central block in the key-developing district. It is the regional economic propeller and transfers the industry development momentum to the rest of the block. These studies provide a static comparison of the evolution and optimization of industrial networks from a dynamic perspective. At the same time, the optimization on the network has a definite direction.
Secondly, from the dynamic perspective, this paper propose a new link prediction model, analyze the evolution characteristics of the regional industrial network structure and identify the path and direction of its structural optimization. Taking the degree centrality in the index of industrial network as an example, the conclusions are: degree centrality which contains wholesale, retail trade and finance has been increasing year by year. The two industries belong to the tertiary industry. After the end of the 12th Five-Year Plan, the degree centrality of the tertiary industry is gradually increasing. The effect of the industrial structure optimization adjustment was fully reflected in the forecasted industrial network. The results of the analysis show that during the 11th Five-Year Plan period, the manufacturing service industry gradually formed and developed continuously. At the end of the 12th Five-Year Plan period, the wholesale and retail trade was not only close to the second industry connected, but also formed the most direct and strong relationship with the primary industry. In addition, in the “Eleventh Five-Year” economic development period, the financial sector is most closely linked to the real estate sector. During the “12th Five-Year” period, the finance has formed the most direct and strong relationship with education, culture, sports, entertainment industry, public management and social organization. This chapter will provide an analytical tool for further research on the optimization of industrial network structure under external shocks.
Thirdly, according to a problem of the regional industrial network structure optimization, this paper constructs a time-series link prediction model with external impact. Then industrial network structure is optimized by adjusting the external impact strength and calculating index. The optimization direction of the network structure is obtained through studying the core-periphery industry and industrial critical path in the network structure.
Finally, industrial symbiosis is an effective measure to realize the optimization of regional industrial network structure. A system dynamics model of regional industrial network symbiosis is constructed based on system dynamics. Based on the analysis of symbiosis optimization between traditional industry and Internet industry, this model considers the activities of industries in the regional industrial networks as a system. Based on the causal relationship between the variables in the regional industrial networks, the model simulates the contribution of industrial symbiosis to the optimization of regional industrial network structure. Through the analysis of policy scenarios, under different scenarios, the paper forecasts the contribution of industrial symbiosis to the regional industrial network GDP, the green decoupling index and ecological contribution to the development of efficiency. In addition, it discerns the main factors that influence the industrial symbiosis in the regional industrial network structure and puts forward the optimization Strategy.
The innovations of this study are as follows:
(1) This paper propose a prediction model of regional industrial network based on link dynamic changing when analyzing the regional industrial network structure and optimizing the path. The new link prediction model considers not only the similarity between the two nodes affected by a common neighbor, and considers but also the varying degree of link weight in the prophase and current network. On the basis of considering the degree of change, this paper further consider the direction of link weight changing, and overcome the shortcomings of similar link prediction model using only the current network to predict.
(2) The paper build a time-series link prediction model with external impact and obtain optimized industry network structure. Then, the strategy of how to optimize the network structure is finally obtained by studying the core-periphery industry and the optimal path of the industry critical link. This paper improves the limitations of optimization analysis from the macro level of network structure in the existing research.
(3) On the basis of the connotation and characteristics of the symbiotic and optimized development of traditional industries and internet industries, the paper analyzes the key variables and causal link between the two industries. The paper construct the corresponding system dynamics model and set the benchmark, yellow, blue and green four target scenarios. Through the simulation analysis of the allocation of policy, the paper finally obtained the optimal policy allocation of the two types of industrial symbiosis. The chapter is the innovation of applied research.