综述了人为差错概率的量化方法。首先，介绍了人为差错概率的基本概念，强调了人为差错概率量化的必要性，指出人为差错概率量化不仅是人机系统概率风险评估的重要组成部分，也是制定人为差错规避措施的重要依据。然后，根据人为差错概率量化理论基础的不同，将现有的人为差错概率量化方法分为时间决定论、任务决定论和场景决定论3种类型，针对3种类型中的典型方法，分别从方法的思路、量化过程以及发展现状等方面进行了详细分析。最后，分别从人为差错数据的使用、场景的表征和使用、量化过程使用的模型以及人为差错概率量化的目标和结果等4个方面，对人为差错概率量化方法进行了综合评述，指出了现有方法中所存在的共性问题，并提出改进建议。
The Human Error Probability (HEP) quantification methods are reviewed. Firstly, the basic concept of HEP is introduced and the necessity for quantify HEP is emphasized. It is pointed that the HEP quantification is not only an important ingredient of Probabilistic Risk Assessment (PRA) for human-machine system, but also can provide a significant guidance for human error reduction. Secondly, considering that there are differences in the theoretical basis of each HEP quantification method, the current HEP quantification methods are categorized into three types; they are time-based, task-based, and context-based. The typical methods of each type are analyzed in details in term of their basic ideas, quantification processes, their developments, and applications. Lastly, the HEP quantification methods are reviewed from aspects of the usage of human error data, the representation and application of scenario, the models used in quantification process, the goal and result of quantification. Some propositions are provided.
