摘要
基于有限Markov链理论和软件统计测试，提出了一种频率计数算法．该算法能够同时解决软件Markov测试中存在的两个基本困难，即建立精确的使用链和确定满足测试充分性的最优用例数．综合考虑了测试执行时间，算例表明了该算法的有效性．
Based on limited Markov Chains theories and statistical software testing , this article presents a kind of frequency count method which can solve two basic difficulties in the Markov software testing simultaneously. Namely： 1. establishing accurate usage chain profiles （i.e. usage probabilities ）; 2. insuring optimal testing times which are sufficient for testing adequacy. The algorithm synthesizes to consider the execution time, and calculates an example to show that method is extremely efficient.
