基于GEP-RNC的指数对数型程序不变量发现方法  

Method of Finding Program Invariants of Exponential and Logarithm Based on GEP-RNC

作　　者：李玉燕[1] 阳小华[1] 吴取劲[1] 

机构地区：[1]南华大学计算机科学与技术学院,湖南衡阳421001

出　　处：《南华大学学报（自然科学版）》2017年第1期72-76,共5页Journal of University of South China：Science and Technology

基　　金：湖南省哲学社会科学基金(14YBA335)

摘　　要：程序不变量的发现是一种提高软件质量的有效方法.不变量发现工具Daikon可以发现程序中蕴含的简单不变量形式,但不包括复杂的函数型不变量.本文基于GEP-RNC算法对指数对数型不变量发现方法进行研究,通过实验证明GEP-RNC算法可以有效的发现指数对数形式的不变量,解决了基因表达式编程算法在复杂函数形式发现中稳定性不佳,精度不高的问题,扩展了Daikon不变量测试库中程序不变量的类型.The discovery of program invariants is an effective method to improve the quality of software.We can find the simple invariants in the program through the test library by u- sing the invariant discovery tool Daikon,but not including the complex function invariants. In this paper,we study the method of finding the invariants of exponential and logarithm type based on GEP-RNC algorithm. It is proved that GEP-RNC algorithm can effectively find the invariants of exponential and logarithmic form, which solves the problems of poor stability and low precision of gene expression programming algorithm in the form of complex function, and extends the type of program invariants in the invariant test library in Daikon.

关 键 词：程序不变量 GEP-RNC 对数 指数 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]