CREAM方法中控制模式分类的模糊方法

认知可靠性与差错分析方法(CREAM)通过对任务环境进行分析从而直接确定人为差错发生概率。该方法提出控制模式的概念,认定不同的任务环境下可能对应不同的控制模式,而人为差错概率直接由控制模式确定。因此,确定控制模式是其中的关键环节。针对该问题,笔者提出了一种模糊方法。通过对性能影响因子进行打分对任务环境进行量化,利用专家打分的方式确定性能影响因子分值对于控制模式的隶属度;通过拟合分析确定性能影响因子的分值对控制模式的隶属函数,得到控制模式与分值之间的对应关系;将该方法应用到两个示例中,其结果与原有方法的结果基本一致。

Determination of Control Modes in CREAM Method Based on Fuzzy Logic

CREAM (cognitive reliability and error analysis method) is used to determine the probability of human errors by analyzing the task context; there raises a concept of control mode in this method that directly determines the probability of human errors. However, different task context determines different control modes. So, to obtain the probability of human error, determining the control mode is the key point. Aiming at this problem, an approach based on fuzzy logic is introduced. First, the task context is quantified by grading the common performance conditions (CPCs); the degree of membership of CPCs to the control modes is determined by using expert judgments. Then, the membership functions, reflecting the relationship between control mode and its score, are acquired by best polynomial approximation. Finally, two examples are illustrated to verify the feasibility and rationality of this approach.