基于贝叶斯方法的通信网络运维人误概率研究

Research on Human Error Probability of Communication Network Operation and Maintenance Based on Bayesian Method

为了定量化分析通信网络运维任务中人因失误的发生概率,针对网络运维人员在执行任务中可能发生的40种人误模式,采用专家判断法获取人误概率的估值,进而采用贝叶斯方法进行聚合分析,得到了各项人误模式的基础发生概率水平。结果表明:在无其他突出情境因素影响时,大多数人误模式的基础发生概率量级在1E-3左右,其中由于理解失效和决策制定失效导致的人误发生概率与其他宏认知功能的失效相比较高。在实际应用中,相对概率高的人误模式在安全和质量管理中应得到优先控制。

To quantitatively investigate the probabilities of various human errors in communication network operation and maintenance tasks,the nominal human error probabilities of 40 common error modes were estimated using expert judgments aggregated by Bayesian methods.The results indicated that the order of magnitudes of most error modes were around 1E-3,given all performance shaping factors at nominal levels.Besides,errors arisen from understanding and decision-making failures were found with higher probability as compared with errors arisen from other cognitive function fail-ures.In practical applications,those error modes with relatively high probabilities should be put greater priority on.