论文部分内容阅读
针对一种计算机网络模型 ,利用节点排队长度累计量的均方涨落函数 ,研究了网络节点在时间上的长程相关特性 .结果表明 ,随着负载的增加 ,网络节点数据包排队长度在时间上由自由流状态的不相关或短程相关逐渐演变为临界和拥塞时的长程相关 ,关联范围逐渐增大 ,长程关联特性开始显现 .在自由流状态时 ,节点的不相关或短程相关 ,并且有一致的数值为 0 .5的幂指数这一典型特征 .而在临界状态时 ,节点数据包排队长度长程相关 ,有大于0 .5的幂指数为特征 .并且随网络规模的增大 ,节点间的群体作用逐渐显著 ,幂指数呈下降趋势
Aiming at a computer network model, the long-range correlation characteristics of network nodes in time are studied by using the mean square fluctuation function of the accumulated length of node queues. The results show that with the increase of load, the queue length of network node data packets in time From the uncorrelated or short-range correlation of free-flow states to the long-range correlations at the critical and congestion levels, the range of correlation increases gradually and the long-range correlation features begin to appear. In the free-flow regime, the nodes are uncorrelated or short-range related and consistent Is a typical exponent of 0 .5 In the critical state, the queue length of node data packets is long-range dependent and has a power exponent greater than 0.5, and as the size of the network increases, The role of the group became more and more obvious, and the power index showed a downward trend