基于Markov理论的多维系统可靠性评价体系构建

The Multidimensional System Reliability Evaluation Framework Based on Markov Theory

摘要: 无线通信系统可靠性是衡量网络性能的重要指标体系，可用于指导用户的个性化网络定制，提升用户的网络体验。由于新一代通信系统中海量数据业务对网络的要求呈现强异质性，网络需要更为全面多元的可靠性评估体系，用以指导网络优化、资源配置。而无线信道的随机性、用户数据生成的偶发性，使得系统可靠性成为一个随机变量。本文基于Markov理论，为无线通信系统构建多维度的可靠性评价体系，将用户数据到达过程、无线链路服务过程分别建模为泊松和指数分布，利用生灭过程描述多信道可用性的随机变化，提出系统丢包率、两次故障发生的平均时间间隔(MTBF)、平均故障时间(MDT)、平均正常运行时间(MUT)四个可靠性评估指标，并分别推导了对应的闭式表达。

Abstract: The reliability of wireless communication systems is an important index to measure network per-formance, which can be used to guide personalized network customization for terminals and im-prove network experience. As the requirements of massive data services in the next generation communication systems show strong heterogeneity, a more comprehensive and diversified reliabil-ity evaluation system is needed to guide network optimization and resource allocation. The ran-domness of wireless channel and the serendipitous generation of user data make system reliability a random variable. In this paper, Markov theory is adopted to construct a multidimensional evalua-tion framework of system reliability. The arrival process of data and the service process of wireless links are modeled as poisson distribution and exponential distribution respectively. The birth and death process is used to describe the randomness of multi-channel availability. The system packet loss rate, the mean time between two failures (MTBF), the mean downtime (MDT), and the mean uptime (MUT) are proposed to evaluate the wireless system reliability. Meanwhile, the correspond-ing closed-forms are derived.