分数混合分数布朗运动关于计算机网络流量模型的一些相关问题-应用统计专业论文.docxVIP

华中 科 技大 华 中 科 技 大 学 硕 士 学 位 论 文 II II ABSTRACT In todays society, the popularity of computer networks and applications permeate all levels of society.however,The limited resource of networks require optimal control.For this purpose it is helpful to use modeling for computer network traffic.To describe mathematical models of these characteristics,it seems reasonable to use a model based on a Poison process or Markov based model.However,in many researchers have been proved that traffic in packet-based network as for example Ethernet or Internet exhibits self-similarity,long-range dependence and complex correlation structures.It is clear that these properties cannot be satisfied by selected previously models.In this situation it is possible to use one of fractal processes,namely fractional mixed fractional Brownian motion. In this paper we study the fractional mixed fractional Brownian motion and some related questions of its application to computer network traffic modeling .Firstly we introduce the knowledge of the development process and the basic idea of the computer network traffic,Brownian motion and fractional Brownian motion and gives the Monte Carlo simulation algorithm steps.secondly through simulate the computer network to generate the simple path of fractional mixed fractional Brownian motion, and constructed a simple path-based programming and draw a fraction of a case of different values fractional mixed fractional Brownian motion images,basing constructor parameter estimation algorithm to construct a simple path in the previous chapter.lastly through Monte Carlo simulation to estimate the parameters,the preparation procedures for parameter estimation, the value of the parameter estimation, and compared with the correct value. The results show that the parameter estimation accuracy is indeed high. We study the fractional mixed fractional Brownian motion’s basic propertied.It can react more real traffic change.Moreover,we developed MFMBM sample paths