Abstract: To fully consider the integrity of index system and various uncertainties of the experts' assessments in the process of selecting key quality characteristics of Computer Numerical Control(CNC)machine tools,a multi-objective comprehensive fuzzy evaluation approach for identifying key quality characteristics of meta action units was developed.A multi-objective quality characteristic index list of CNC machine tools was constructed from the meta action unit perspective,and the related concepts and definitions were given.Then,rough number theory was utilized to transform the quality characteristics evaluations into interval form to cope with the vagueness and subjectivity of the experts' judgments.The rough interval evaluations were further converted into the normal cloud form to address the problem that the rough number theory was not enough to reflect the randomness of experts' evaluations.Combined with the weight information of each expert,the relative weights of all evaluation factors were computed by using the rough order weighted average method.Aiming at the issue of lacking the mechanism of manipulating the experts' bounded rationality in conventional decision-making method,an evaluation method considering risk preference was employed to rank the quality characteristics and identify the key quality characteristics.The effectiveness of the developed method was illustrated by an application case.

Key words: key quality characteristics, meta action unit, rough number theory, cloud model theory, risk preference decision-making, computer numerical control machine tools

摘要： 为了充分考虑数控机床关键质量特性识别过程中指标体系的完整性和评估信息的多种不确定性,提出一种面向数控机床元动作单元的多目标综合模糊评价关键质量特性识别方法。从元动作单元角度构建数控机床多目标质量特性指标体系,并给出了相关概念及定义。采用粗糙数理论将质量特性评估信息转化为区间形式去处理评估信息中的模糊性和主观性,再将粗糙区间评价值转化为正态云形式,以弥补粗糙数不能处理信息随机性的缺陷。结合专家权重的分析结果,采用粗有序加权平均法求解评价因子的相对重要性。针对传统决策方法没有考虑专家有限理性的问题,使用基于风险偏好的评估方法进行质量特性排序和关键质量特性的识别。最后,通过实例验证了所提方法的有效性。

关键词: 关键质量特性识别, 元动作单元, 粗糙数理论, 云模型, 风险偏好决策, 数控机床