Automatic Testing Data Generation in the Testing Based on EFSM

张涌,钱乐秋,王渊峰

DOI: https://doi.org/10.3321/j.issn:0254-4164.2003.10.012

2003-01-01

Jisuanji Xuebao/Chinese Journal of Computers

Abstract:We propose an automatic testing data generation method in the testing based on EFSM. In this method, we use two key phases, interval narrowing and subsection gradient optimal descent algorithms to generate testing data automatically. In the interval-narrowing phase, we design a group of interval narrowing operators to reduce the variables' value scope. If the precondition in the state transitions is simple, we can get the testing data in this phase. In the secondphase, we use subsection gradient optimal descent algorithms to generate testing data from reduced variables' value scope. The experiment shows that we can generate most of testing data automatically using this method, and has speedy convergence. In the situation of having no solution of input variables, interval narrowing can provide small intervals of the value of input variables for tester, which would be convenient for tester to select testing data from those small intervals.