首页 | 本学科首页   官方微博 | 高级检索  
     

闭排队网络当性能函数与参数相关时的性能灵敏度分析
引用本文:殷保群,周亚平,奚宏生,孙德敏.闭排队网络当性能函数与参数相关时的性能灵敏度分析[J].控制理论与应用,2002,19(2):311-312.
作者姓名:殷保群  周亚平  奚宏生  孙德敏
作者单位:中国科学技术大学自动化系,合肥,230026
基金项目:国家自然科学基金(69974037); 国家高性能计算基金(00212)资助项目.
摘    要:借助于无穷小矩阵摄动方法, 讨论了一类Markov过程, 其稳态性能关于参数摄动的灵敏度分析问题. 然后研究了闭排队网络的稳态性能灵敏度分析问题, 并在参数相关性能函数的情况下, 给出了网络的几种稳态性能的灵敏度公式. 这些公式表明稳态性能灵敏度很容易通过网络势能进行计算.

关 键 词:闭排队网络    参数摄动    灵敏度分析    势能
文章编号:1000-8152(2002)02-0311-02
收稿时间:6/9/1999 12:00:00 AM
修稿时间:3/2/2001 12:00:00 AM

Sensitivity Analysis of Performance with Parameter-Dependent Performance Functions in Closed Queueing Networks
YIN Baoqun,ZHOU Yaping,XI Hongsheng and SUN Demin.Sensitivity Analysis of Performance with Parameter-Dependent Performance Functions in Closed Queueing Networks[J].Control Theory & Applications,2002,19(2):311-312.
Authors:YIN Baoqun  ZHOU Yaping  XI Hongsheng and SUN Demin
Affiliation:Department of Automation, China University of Science and Technology, Hefei, 230026, P.R.China;Department of Automation, China University of Science and Technology, Hefei, 230026, P.R.China;Department of Automation, China University of Science and Technology, Hefei, 230026, P.R.China;Department of Automation, China University of Science and Technology, Hefei, 230026, P.R.China
Abstract:By using the approach of the infinitesimal generator perturbation, the problems of sensitivity analysis of the steady-state performance with respect to the parameter perturbation are discussed for a class of Markov processes. Then the sensitivities of steady-state performance are studied in closed state_dependent queueing networks. Sensitivity formulas of several kinds of the steady-state performance are given for parameter-dependent performance functions. These formulas show that the sensitivities of steady-state performance can be easily calculated by the potentials of networks.
Keywords:closed queueing networks  parameter perturbation  sensitivity analysis  potential
本文献已被 CNKI 维普 万方数据 等数据库收录!
点击此处可从《控制理论与应用》浏览原始摘要信息
点击此处可从《控制理论与应用》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号