In The Complex System Simulation System Evolution Complexity Research Of Discriminant Method

From the end of the last century to the beginning of this century，The simulation ofthe complex economic system boomed and developed in the research of complex systemstheory which is As the representative of the CAS theory, and emerged a group of valuableresearch in exploratory and application. Researchers of Santa Fe Institute (SFI) introducedsimulation method into complexity researching, developed the artificial stock marketmodel, and Sandia National Laboratory developed the U.S. economic model ASPEN withintelligent agents. However, in recent years, they gradually fall in trouble, due to thelacking of systemic theoretical guidance In the research of this field,besides the reasonthat, lacking of new breakthrough in the theories of complexity science make thesimulation of complex economic system lost in the forward direction and power, Moreimportant reason is that CAS theory uses the metaphor theory to represent the complexityand is lacking of a strict mathematical definition in complexity,and generally usemorphology discrimination which identified the phenomena and laws of the results of themodel output complexity with the naked eye. But, the way people use to be able to identifythe complexity is very limited, and often with subjective color.My article hope to analyze the simulation results of the simulation model byintroduced chaos discrimination method, so that can expand and explore new avenues inthe simulation study of the complex economic system. Discrimination methods of chaotictime series have power spectrum method, principal component analysis (PCA distribution),the Poincare section method, Lyapunov exponent, correlation dimension, instead of thedata method, and so on. This discriminant method which the article used is based on phasespace reconstruction, Get the correlation dimension, through calculating the correlationintegral by using G-P algorithm, and to distinguish chaotic time series with the increasingtrend of correlation dimension.In this study, I will use the imitate ASPEN model which was developed by my Mentorin the projects of National Natural Science Foundation, as research material, throughsimulation experiments and data analysis to determine the feasibility of chaotic time seriesanalysis methods in the discrimination the complexity related to the evolution, andconfirmed the validity of this approach. The study will focus on economic simulationmodel output data exists Chaos, and make this as the basis of discriminating it existsevolution complexity, In the model output data, The vendor profit targets in imitation of the ASPEN model has the typical characteristics of the aperiodic fluctuations, and theindicator is directly related to the game process of the specific type of vendor productpricing. So in this study I analyzed the characteristics of the food vendors. Use of chaotictime series analysis methods, to research the model output data time series of the subject.