实时系统多路径任务概率时序分析研究综述  

Survey of Probabilistic Timing Analysis of Multipath Tasks for Real-time System

作　　者：李子豪 张轶[2] 刘学 邓庆绪[1] LI Zihao;ZHANG Yi;LIU Xue;DENG Qingxu(School of Computer Science and Engineering,Northeastern University,Shenyang 110004,China;College of Medicineand Biological Information Engineering,Northeastern University,Shenyang 110004,China)

机构地区：[1]东北大学计算机科学与工程学院,沈阳110004 [2]东北大学医学与生物信息工程学院,沈阳110004

出　　处：《小型微型计算机系统》2024年第11期2586-2593,共8页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(62072085)资助.

摘　　要：随着现代嵌入式实时系统在硬件与软件结构上的日益复杂化,传统WCET分析方法已难以满足需求,这推动了使用概率和统计技术对实时系统中多路径任务进行时序分析.本文深入探讨了这一领域的研究成果,首先回顾了传统时序分析方法,总结了传统方法的实现步骤和局限性,随后引入了概率时序分析和统计学的基本概念.接着详细讨论了静态概率时序分析、基于测量的概率时序分析和混合概率时序分析3种技术路线,并重点分析了这些技术在多路径任务时序分析中的应用与发展现状.最后,总结了当前多路径任务概率时序分析技术存在的主要挑战,并对未来的研究方向提出了建议.With the increasing complexity of modern embedded real-time systems in both hardware and software structures,traditional Worst-Case Execution Time(WCET)analysis methods have become inadequate,leading to the utilization of probability and statistical techniques for timing analysis of multi-path tasks in real-time systems.This paper delves into the research outcomes in this domain,starting with a review of traditional timing analysis methods,outlining the implementation steps and limitations of these conventional approaches.Subsequently,it introduces the fundamental concepts of probabilistic timing analysis and statistics.The paper then extensively discusses three technical pathways:static probabilistic timing analysis,measurement-based probabilistic timing analysis,and hybrid probabilistic timing analysis,focusing on the applications and current developments of these techniques in the timing analysis of multi-path tasks.Finally,it summarizes the primary challenges existing in current probabilistic timing analysis of multi-path tasks and provides recommendations for future research directions.

关 键 词：多路径 概率时序分析 实时 WCET 

分 类 号：TP316[自动化与计算机技术—计算机软件与理论]