多线程程序时序分析的隐Markov模型.pdfVIP

ISSN 1000-9825, CODEN RUXUEW E-mail: jos@ Journal of Software , Vol.21, No.3, March 2010, pp.461−472 doi: 10.3724/SP.J.1001.2010.03521 Tel/Fax: +86-10 © by Institute of Software , the Chinese Academy of Sciences . All rights reserved. ∗ 多线程程序时序分析的隐 Markov 模型 + 孔德光 , 谭小彬, 奚宏生, 帅建梅, 宫 涛 ( 中国科学技术大学 自动化系,安徽 合肥 230027) Hidden Markov Model for Multi-Thread Programs Time Sequence Analysis + KONG De-Guang , TAN Xiao-Bin, XI Hong-Sheng, SHUAI Jian-Mei, GONG Tao (Department of Automation, University of Science and Technology of China, Hefei 230027, China) + Corresponding author: E-mail: kdg@ Kong DG, Tan XB, Xi HS, Shuai JM, Gong T. Hidden Markov model for multi-thread programs time sequence analysis. Journal of Software, 2010,21(3):461−472. /1000-9825/3521.htm Abstract : To overcome the difficulty of analyzing and detecting the data race in multithread programs, a method based on Hidden Markov model is presented for the analysis of time sequences in multithread programs. The random variable uncertainty is used to depict the mutual influence in different multithread in time sequences, and the probability distribution for random variable uncertainty is analyzed as the outcome of multithread programs on the condition of data race. Hidden Markov model is constructed to appreciate the state for the thread running according to the observed values of the running threads. The Baum-Welch and forwarding algorithm are used to simulate the real running process of the programs in the influence of context. It