Complex System Modeling Methods And Investment Analysis

With the rapid development of scientific theories and production technologies, continuous innovations are also happening to modern management theories. The Theory of Complex System is an important part of System Science, and it is inevitable to greatly influence the development of System Science.The problems facing us are complex and non-linear. Many variables in economic and political systems are non-linear, and the linear method is out-dated. Although the development of Complex System has just begun, it has been called by some scientists "the science of the 21st century". Hawking, a famous Britain theoretical physicist, once said, "I think the next century will be a century of complexity." This means that the task of the theoretical science in 21st century is to deal with Complex System, and construct a theory system which is different from before.Complex System is multidisciplinary. It is related to mathematics, physics, non-linear science, management and so on. It is being applied to many subjects, such as finance management, and has become a multidisciplinary area that is concerned by many scientists. The combination of Complex System and finance management in capital markets is an important part of the research of Complex System. This article talks about the application of Complex System in investment analysis, both qualitatively and quantitatively.This article summarizes the existing theories and modeling methods of Complex System; analyzes the Herd Behavior in stock market with Complex Network, probes into some macro behaviors in stock market, and concludes that the Herd Behavior has some characteristics of Scale-free Network; analyzes the capital market with Fractal Theory, carries out investment forecasting with Hausdorff dimension and varied fractal dimensions, and talks about the limitations of Fractal Theory; solves the non-linear portfolio question with Genetic Algorithm, and programs with MATLAB5.3 and Sheffield GA toolbox; compares the above modeling methods in terms of scale-free, preferential attachment, and growth.