基于GEP的线性指数函数型程序不变量动态发现方法

不变量是用来描述程序运行时保持不变性质的逻辑断言.根据关系数据理论,程序不变量可分函数依赖型和非函数依赖型程序不变量.着眼于函数依赖型程序不变量,借助GEP的函数发现特点和Daikon对线性程序不变量的发现能力,重点对线性指数函数型程序不变量动态发现方法进行研究,通过实验证明了GEP对线性指数形式的函数有较高的发现效率,可以扩展Daikon在线性指数函数型程序不变量方面的预置形式以达到从程序轨迹数据中发现该类程序不变量的目的.

Methods on Lineal and Exponential Function Program Invariants Dynamically Mining Based on GEP

Invariants is used to describe logical assertions of which running programs keep unchanged.According to the relational data theory,program invariants can be divided into Function Program Invariants and Non-Function Program Invariants.The thesis beginning from function program invariants focuses on the research on Lineal and Exponential Function Invariants Dynamically Mining with the help of GEP feature of Function Mining and Daikon ability to find out the Lineal and Exponential Function Program Invariants.The experiment proves the high efficiency of GEP discovering Lineal and Exponential function type,which expand the type Daikon preset for Function Program Invariants in order to find more invariants of this type from Program-Running track data.