Research Of Chaos Theory Is Used To Predict The Network Game Flow

In recent years due to the rapid development of computer hardware technology and network communication technology, the electronic game industry has become more and more mature, is one of the fastest growing online game industry, huge network market makes the development of the network game was caught the attention of the society from all walks of life, especially a growing number of IT companies and game developers start into the network game industry, network game players are increasingly at the same time, then there is the network game server had a lot of pressure on the performance and load. If you want to give the network game players to provide the best service and reduce the operation and maintenance cost is the key to control the server architecture and load balancing.This paper chose to network game flow prediction as the research object, to study the fractal characteristics of network traffic, in this paper, the main research contents include: first, to analyze the network traffic data, to the phase space reconstruction of traffic data, calculate the correlation, so the fractal characteristic of time series analysis flow, realize the determination of time delay, autocorrelation function, the initial network flow characteristics of chaos. And then reconstructs the phase space, determine the minimum embedded peacekeeping network traffic time-series macro- the correlation parameters, so as to realize the quantitative description of network traffic behavior characteristics. Once again to verify the dynamic behavior of network traffic information system is chaotic and its attractor of fractal structure. Based on network traffic behavior has fractal characteristics, based on the phase space reconstruction theory, established a chaotic model, used in network traffic prediction; And then to establish the model of simulation and verification; Verification results show that the predicted model and the real value of relative error between 0.02 to 0.08; After using the model of the author’s group a game server for a game company monitoring, prediction accuracy is between 0.02 to 0.11.In this paper, we studied the fractal characteristic of the network traffic and the related design and practice of forecasting model analysis, to understand and solve the network game flow, be load balancing server group and configuration has better practical reference value.