The Research Of Industrial Networks Based On Complex Networks Theory

With deepening and expand of the global economic division, networknization between enterprises is more and more obvious; in that case, studying the relationship within the industry and between the industry and their influence on the economic subject from the perspective of network, is becoming the important research topics for national economic system gradually. In fact national economic system can be described by the system which includes the production departments and consumption departments with different natures and mutual relationship, which formed an association networks of industrial departments, that called industrial networks for short.This paper is based on the complex networks theory, makes each industry as a point and the relationship(input-output) between industres as an edge, then we establish a weighted and directed networks, and with the focus on the complex characteristics and industrial networks dynamic evolution etc.launches out nto research work.First of all, this paper gives a simple review on the basic concept of input-output, and a summary and analysis about the current industry networks study status, and then presents industrial networks model based on the input-output; Meanwhile, the current study content is summarized about complex networks, which provides theoretical support for this paper.Secondly, this paper is to establish a industrial networks model based on the product magnitude of value, after that a industrial networks is established on published department of national economic accounting. Then the paper gives a characteristic empirical analysis for industrial networks, to study every industry operating efficiency and so on.Thirdly, with random, degree value standard and betweenness value standard to attack the industrial networks, we can have an anti-destroying ability analysis and research, and put forward improvement strategies of the industrial networks to raise the entire the operating efficiency of industry system.Fourthly based on the weighted and directed characteristics, this paper put forward an improved K-means clustering algorithm, and with combining modularity industrial networks is partitioned into five communities, and established communities relational matrix to analysis the input-output relationships between communities.Finally, this paper attempts to analyze complex characteristics with input-output tables of china in 2007,2002 and 1997, find out new contents and evolution regularity of industry, and put forward the rule model of dynamic networks characteristic.