Study Of Formation Mechanism Of Industry Cluster

We attempt to give a summary for diverse theories of industry clusters and point out that there, in fact, is still not a satisfactory model of industry clusters. So far, the extensively acceptable model of industry clusters is core-periphery model of Krugman. However, more and more economists started to be not satisfied with this model since it neglected correlations among firms. In this paper, we attempt to set up a model of industry clusters involving correlations among firms. Our study shows that the correlations among firms can be described by complex networks and that the fractal dimension of these complex networks can be regarded as a characteristic variable. By taking advantage of the ways of nonlinear dynamical systems and statistical physics, we can get an evolutive equation of firms involving fractal dimension. This equation shows that, if we neglect the correlations among firms and suppose homogeneous geography, we shall not get industry clusters and conclude industry diffusion. Our equation is based on macroscopical viewpoint and gives the distribution of firms in geography, it uncovers the fractal structure of industry clusters. The cost of transportation or endowment is one of all sufficient conditions for forming industry clusters. As the most important index, the fractal dimension will show the hierarchical structure of economic system. The relational network of industry clusters that has larger network density may be chaos, namely, the fractal dimension approaches zero. The chaos will induce the benefit of industry clusters decline, even, callapse. Although measure of fractal dimension of relational network is an unresolved question, we still attempt to suggest a way of measure. Finally, we give the relevant policy proposal.