时钟有限自动机模型及其演化算法  

Clock finite state automata model and its evolution algorithm

作　　者：范林军 史湘宁[2] 凌云翔[1,3] 

机构地区：[1]武警警官学院管理科学与工程系,四川成都610213 [2]湖南商学院计算机与信息工程学院,湖南长沙410205 [3]国防科学技术大学信息系统工程重点实验室,湖南长沙410073

出　　处：《计算机工程与科学》2017年第2期378-384,共7页Computer Engineering & Science

基　　金：国家自然科学基金(61272336)

摘　　要：传统分布仿真系统时钟不一致影响因素分析方法,已不能满足当前面向服务分布仿真的时钟状态分析需要。从系统全局时钟演化出发,阐述了时钟状态演化内涵与过程;在此基础上,基于有限自动机理论,提出了用于时钟不一致影响因素量化分析的动态演化模型及其算法:时钟有限自动机CFSA和时钟一致性演化算法CCEA。仿真实验表明:相比传统的分析方法,使用CFSA模型及其CCEA演化算法刻画系统时钟一致性状态变迁过程,探寻各种不一致因素的影响机理,量化分析各因素的影响程度等具有可行性、有效性和新颖性,可为面向服务分布仿真中时钟同步算法设计提供指导性建议。Given that the analysis methods of clock inconsistency factors used in traditional distributed simulation systems cannot meet the analysis requirements of clock states in current service-oriented distributed simulation applications, we explore the evolution mechanism and transition processes of system clock states from the perspective of global clocks, and propose a new dynamic evolution model named clock finite state automata （CFSA） based on the finite state automata theory. We then design an evolution algorithm named clock consistency evolution algorithm （CCEA） in order to quantitatively ana- lyze clock inconsistency factors. The results of simulation experiments show that compared with traditional analysis methods, it is feasible, effective and novel for using the CFSA and CCEA to characterize the transition processes of system clock states, to investigate the influence mechanism of various inconsistency factors and to quantitatively analyze the impact of various factors, thus providing guidance and advice for the design of clock synchronization algorithms in distributed simulation applications.

关 键 词：计算机应用 分布式系统 时钟一致性演化 因素量化分析 有限自动机 形式化方法 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]