基于确定有限状态机的测试输入序列选取

TEST SEQUENCES SELECTION BASED ON DETERMINISTIC FINITE-STATE MACHINES

有限状态机可以精确地刻画软件系统或其子系统的行为 ,其在软件建模中被广泛应用 .基于有限状态机的测试问题得到了广泛的研究 ,其中 Wp方法由于限制条件较少 ,并且其可以达到较高的错误覆盖度 ,因此被广泛使用 .但当有限状态机的实现中的状态数目的估计值 m较大时 ,产生的测试输入序列数目过多 ,影响了其使用效率 .提出一种 Wp方法的改进方法 R- Wp方法 ,使用该方法在 m值较大时可以产生相对较少的测试输入序列数目 ,可以提高测试效率 ;同时本文对 Wp方法和 R- Wp方法产生的测试输入序列数目进行了讨论 ,并证明了该方法与Wp方法相同的错误覆盖能力 .

The behavior of a software system or its subsystems can be described precisely using finite-state machines, so it has been used widely in software modeling. Many researchers have proposed some test selection methods based on FSM. Among these methods, the Wp-method, which has fewer use constraints and higher fault coverage, can be used more widely. But if the estimated states number ( m ) of the implementation of a specification is large, a large number of test sequences will be generated using the Wp-method, which will decrease the testing efficiency. In this paper, an improved method of the Wp-method, named R-Wp method, is presented. When the m is large, fewer test sequences can be obtained through this method than through the Wp-method. In addition, the number of test sequences generated from the R-Wp method and the Wp-method, and factors which influence the number of test sequences generation, are discussed. Finally, it is proved that the R-Wp has the same fault detecting ability as the Wp-method.