摘要
运用排队理论对视频会议系统中的音频混合问题进行了分析.假定两类音频包的到达为相互独立的泊松过程,到达混合模块后分别在各自的有限容量缓冲区中排队,混合器的服务规则依据队列排队状态而定,服务时间服从负指数分布.对混合模块建立了具有两个队列的M/M/1排队模型,给出状态转移图和相应的Q矩阵,并采用矩阵分析的方法对模型求解,得出了两类音频包各自的稳态分布.最后作了相应的性能分析,给出了平均队长,溢出概率等的计算公式.
Queueing theory is applied in the analysis of voice mixing of video conference. The arrival processes of the two kinds of voice packets are all Poisson processes. After arriving at the mixing module, each packet queues in its finite capacity buffer. The mixer serves the packets according to the state of each queue. And the service rate is exponential. A M/M/1 queueing model with two queues is formed. The state changing figure and Q matrix are given. Steady-state distribution is worked out and some performance evaluations, such as average queue length and overflow probability are given by using matrix analysis.
出处
《江苏大学学报（自然科学版）》
EI
CAS
北大核心
2006年第2期181-184,共4页
Journal of Jiangsu University：Natural Science Edition
基金
国家自然科学基金资助项目(70571030)
关键词
排队论
视频会议
音频混合
有限容量缓冲器
稳态分布
queueing theory
video conference
speech mixing
finite capacity buffer
steady distribution
