数据并行传输程序中代码坏味检测算法研究  

Research on Code Bad Smell Detection Algorithm in Data Parallel Transmission Program

作　　者：于萧榕[1] 黄健荣[1] YU Xiao-rong;HUANG Jian-rong(College of Computer Science,Jiangsu University of Science and Technology,Zhejiang Jiangsu 212003,China)

机构地区：[1]江苏科技大学计算机学院,江苏镇江212003

出　　处：《计算机仿真》2021年第5期403-407,共5页Computer Simulation

摘　　要：程序中某些代码坏味会导致整个程序无法顺利运行,此时,需要清除引起程序错误的代码坏味,以保证程序的正常运行。为此提出了一种基于数据并行传输程序的代码异味检测算法。在并行数据传输程序的信道上建立两组独立的数据序列,通过抑制载波幅度调制将它们调制成一对正交载波,由每个子信道中发送的单个脉冲响应决定信道失真情况;利用贝叶斯网络将最新的检测规则转化为概率模型,并从源代码中提取代码坏味度量,实现代码坏味检测。实验结果表明,所提方法的检测结果比Jaccard测距法更准确,保证了整个数据并行传输程序的稳定运行。The bad smell of some code in the program will cause the whole program not to run smoothly. In order to ensure the normal operation of the program, we need to eliminate these codes. Therefore, a code odor detection algorithm based on data parallel transmission program is proposed in this work. Firstly, on the channel of parallel data transmission program, two independent data sequences were constructed. Secondly, based on the amplitude adjustment of the suppressed carrier, the data sequence was modulated into a pair of orthogonal carriers. Then, the distortion of the channel was determined by the response of a single pulse sent in each subchannel. Eventually, according to Bayesian network, the latest detection rules were transformed into probability model, and the bad smell measurement of code was extracted from the source code, thus achieving the detection of bad smell of code. The results show that the detection result of this method is superior to that of Jaccard ranging method, which can ensure the stable operation of the whole data parallel transmission program.

关 键 词：数据并行传输程序 子信道 代码坏味 检测算法 贝叶斯网络 

分 类 号：TP314[自动化与计算机技术—计算机软件与理论]