通过分析如何将概率统计的思想应用到基于有限状态机的协议被动测试上,在Markov链模型基础上,提出了一种新的被动测试错误检测算法,并给出了与已有算法的比较.从比较结果可知,新算法只需要知道有限状态机中转换的概率分布和最终观察到的输入/输出对的概率分布,就可以解决已有算法存在的问题,因此适用范围更广,同时还探讨了单个错误定位问题.
A method for applying statistical observations in passive testing based on finite state machine (FSM) was introduced. Based on the Markov chain model, a new fault detection algorithm was proposed and a single fault diagnostic was discussed. Compared with those existed algorithms, the new approach only requires knowledge of the probability distribution of the transitions and the observed input/output pairs, and it can potentially be used in testing/monitoring distributed systems.