在线评测系统中的源码相似度检测研究与实现

Study on similarity detection of source code in online judge system

在线评测(online judge,OJ)系统是面向程序设计的源码自动评判系统,目前,许多高校将其扩展并应用到程序设计类课程的辅助教学中。如何有效地检测出源码的抄袭是其有效实施的关键问题之一。结合属性计数算法和Rabin-Karp算法对源码进行比对,提出了一种适合于OJ系统源码检测的方法,能自动地对源码进行词法和语法分析,并在比对过程中缓存相关预处理数据,进一步提高检测效率。实验结果表明:该检测方法能有效地检测出OJ系统中的源码抄袭,同时在台州学院在线程序设计综合实验平台中运行效果良好。

The online judge system(OJ)is used to automatically judge the source code for programming and is used as teaching aids for some programming courses.How to effectively detect the source code of plagiarism is one of the key problems in teaching.A source code detection method is proposed for the OJ system.It automatically performs the lexical and grammatical analysis,and compares the source codes by combining attribute counting and Rabin-Karp algorithms.Meanwhile,the preprocessing data is cached to improve the efficiency for next detection.The experimental results show that our method is well suitable for the OJ system such as that in Taizhou University Online Comprehensive Programming Laboratory.