Abstract: A reliability allocation method that considered both cognitive uncertainty and random uncertainty was proposed to address the problems of uncertainty in expert evaluation in meta-action reliability allocation. The mechanical products were reasonably decomposed by using the hierarchical method of function-motion-action(FMA) to obtain the basic motion type(meta-action) of the products. The experts were grouped and evaluated by probabilistic linguistic term set(PLTS) to reduce the random uncertainty of the experts evaluations. The combined evaluation results of the experts in each group were used as the evidences for the inference. And according to the evidence fusion theory to synthesize all experts information and reduce the influences of experts cognitive uncertainty. The weight value of each meta-action was calculated, and then the product reliability index was assigned to the meta-action. The effectiveness and practicality of the method were demonstrated through application examples.

Key words: mechanical product, cognitive uncertainty, random uncertainty, reliability allocation, meta-action

摘要： 针对元动作可靠性分配中专家评价的不确定性问题，提出了一种同时考虑认知不确定性和随机不确定性的可靠性分配方法。利用功能运动动作（FMA）的层次方法将机械产品进行合理分解，得到产品的基本运动形式（元动作）；将专家进行分组并以概率语言术语集（PLTS）进行评价，以降低专家评价的随机不确定性；将各组专家综合评价结果作为证据进行推理，根据证据融合理论综合所有专家信息，并减小专家的认知不确定性影响；计算各元动作权重值，进而将产品可靠性指标分配给元动作。通过应用实例，验证了所提方法的有效性和实用性。

关键词: 机械产品, 认知不确定性, 随机不确定性, 可靠性分配, 元动作