最大熵原理在网络延迟测量数据处理中的应用

Application of maximum entropy in the network delay measurement data processing

在计算机软件时钟同步系统中,针对网络延迟测量数据处理中掺杂的主观因素不能准确反映客观现实的问题。采用最大熵方法,根据已有的测量数据求取被测量的概率分布,进而对此概率分布在约束条件下进行估计和评价。测量数据源自网络延迟测量产生的对数正态分布测量数据样本,计算测量结果用MATLAB编程做出仿真图形。仿真结果与计算结果表明采用最大熵方法所确定的概率分布是含有最少主观假定的分布,取得的被测量的概率分布更加接近真实分布。

Clock synchronization in computer software systems,data processing for network delay measurement of subjective factors in the doping does not accurately reflect the objective reality problem.Using the maximum entropy method,based on the existing measurement data to strike a probability distribution to be measured,then this probability distribution is estimated under the constraints and evaluation.The measurement data came from the network delay measurement sample data with log-norma1 distribution.The result is validated by the MATLAB software.Simulation and calculation results prove that the maximum entropy probability distribution is determined with the least subjective assemened distribution,obtained by measuring by measuring the probability diatribution of a more realistic diatibution.