Reliability allocation method combined with real working conditions under rough probabilistic linguistic environment
Yuan Zhong,Guofa Li,Jialong He,Chuanhai Chen,Yan Liu
DOI: https://doi.org/10.1016/j.eswa.2023.119964
IF: 8.5
2023-04-07
Expert Systems with Applications
Abstract:Reliability allocation is an important part of product design. Whether the result of reliability allocation is reasonable or not is directly related to the ability of the selected subsystem to meet design requirements. This paper proposes a new reliability allocation method to overcome the limitations of traditional reliability allocation methods. These limitations include weak ability to express and process fuzzy information, unreasonable use of data and lack of integration with real working conditions of products. Firstly, rough probabilistic linguistic term set (RPLTS) is proposed by combining rough set theory (RST) and probabilistic linguistic term set (PLTS). Moreover, RPLTS is extended to Muirhead mean (MM) operator and Best-Worst method (BWM), rough probabilistic linguistic weighted MM (RPLWMM) operator and rough probabilistic linguistic BWM (RPLBWM) are proposed to improve the ability of experts to express and process fuzzy information. Secondly, the aggregation results of the RPLWMM operator are only used as subsystem ranking, which not only improves the rationality of data use but also provides the basis for experts to establish the judgment matrix, which reduces the influence of experts' subjective factors. In particular, this paper introduces the proportional hazards (COX) model into reliability allocation for the first time, combining the reliability allocation results of key subsystems with the real working conditions and improving the pertinence and accuracy of reliability allocation. Finally, the advantages of the proposed method are illustrated by a numerical control machine tool case analysis.
computer science, artificial intelligence,engineering, electrical & electronic,operations research & management science