Q Learning-Based Dynamic Dependence Analysis of Complex Programs
Wang Yubao 1,2 Shi Liang 1,2 Xu Baowen 1,2,3,4 (Department of Computer Science and Engineering, Southeast University 1), Nanjing 210096) (Jiangsu Institute of Software Quality 2), Nanjing 210096) (State Key Laboratory of Software Engineering, Wuhan University 3), Wuhan 430072) (Department of Computer, National University of Defense Technology 4), Changsha 410073)
Accurately describing dynamic characteristics of complex programs in indeterminate environment is a challenge to traditional theory and technology of program analysis. And it is also a crucial problem in real-time analysis and control of dynamic programs for many important systems. Based on Q-learning method, the paper presents a novel idea of program dynamic dependence analysis, and gives an algorithm which can make analysis for complex programs more concise, more intelligent and more efficient.