Internet网络访问直径的短期及长期预测

Short-term and Long-term Forecast of Internet Traveling Diameter

本文首先形式化定义表征Internet的参量—访问直径,然后根据海量数据变化特点,提出了基于Logistic模型的、以正余弦函数模拟振荡变化的数学模型,使用浮点型遗传算法确定模型形式。由于Internet访问直径演化复杂,传统一维模型在长期预测中已不适用。因此在长期预测方面,首先计算得到Internet访问直径的关联维数,然后根据关联维数及奇异吸引子相近空间混沌轨道运动特性,提出了基于三维常微分方程组数学模型形式。

Based on giant samples, a property of Internet-traveling diameter-was firstly defined, and its mathematical model which was composed by a Logistic Model and sine and cosine function groups simulating oscillations during the growth of traveling diameter was then determined by a Float-point GA. Since growth of traveling diameter is complex, the model mentioned above is not suitable while being used for a long-term forecast. To solve this problem,Correlation Dimension of traveling diameter was calculated, and a model comprising three-dimension function groups was put foreward based on the value of Correlation Dimension and properties of track near the strange attractor in Chaos system. This model was proved to be comparatively suitable for long-term forecast.