Research on New Methods for Performance Evaluation of Communication Networks

14名成员致力于研究通信网络性能评估的新方法。我们每月举行一次会议，每年举行一次研讨会，参加人数超过100人。输入过程的研究：我们主要研究了长相关和多输入队列。一个显着的结果是：如果我们只考虑损失概率，输入过程在高峰期的行为是最重要的，我们不需要关心输入过程是否是长程相关的.极端情况研究：主要研究了客户分布尾部的几何衰减性质。利用矩阵几何方法或大偏差理论，证明了在许多情况下，尾部是几何衰减的.拥塞传播问题研究：该问题的研究比较困难，目前几乎没有关于拥塞传播的研究结果.网络控制的研究：我们分析了各种网络，并提出了一些关于网络配置和控制的建议。这些结果将在不久的将来发表在期刊上。我们开始了对移动的通信系统和费用控制的研究。这些模型成为通信网络研究的重要课题。

Fourteen members worked on the study of new methods for performance evaluation to communication networks. We had monthly meetings and annual seminars to which more than a hundred people attained.We forcused on the following four topics :1. Study of input processes : We mainly worked on long-range dependence and on multiple-input queues. A remarkable result is the following : If we concerned only loss probabilities, the behavior of the input process in the peak period is most important and we need not concern about whether the input process is long-range-dependent or not.2. Study on extreme situation : We mainly studied geometric decay property of the tail of the customer distribution. Using the matrix-geometric approach or the large-deviation theory, we proved the tail decays geometrically in many cases.3. Study on propagation of congestion : This problem was vary hard and we obtained almost no result about the phenomena.4. Study on control of networks : We analyzed various networks and made a number of proposals on the network configurations and controls. These results will be published in journals near future. We started studies on mobile communication systems and on control by the fee. These mold become important topics in the study of communication networks.