Scholarlink

<p>Selecting appropriate multi-criteria decision-making (MCDM) methods and assigning suitable weights are critical for deriving robust results from Pareto frontier data. To address this, our study systematically analyzes the weight sensitivity of 15 MCDM methods, focusing on their application in multidisciplinary design optimization (MDO) for two automotive structures. We first introduce the concept of critical weight and investigate its underlying mechanisms, revealing that MCDM methods can be classified into two distinct categories: linear methods (e.g., WSM), which exhibit significant sensitivity to critical weights under uniform distributions or strongly conflicting objectives, and nonlinear methods (e.g., VIKOR), which demonstrate greater robustness due to their nonlinear aggregation mechanisms, etc. Based on these findings, we propose a robust decision-making framework that integrates analysis of objective distribution uniformity, conflict quantification, critical weight calculation, and robustness validation. When applied to the MDO of a battery pack system, this framework reduces the decision failure rate to 9.1%, significantly improving stability. Our study thus provides a data-driven methodology for selecting and weighting MCDM methods in automotive MDO, enhancing both reliability and practicality.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Classification, selection of MCDM methods and robust decision-making in multidisciplinary design optimization of automotive structures

  • 期刊论文

  • 发表日期：

    2025年10月30日

查看全文
  • Lian Fengmin,
  • Wang Dengfeng,
  • Xie Ying,
  • Meng Zihao,
  • Chen Jing

摘要

Selecting appropriate multi-criteria decision-making (MCDM) methods and assigning suitable weights are critical for deriving robust results from Pareto frontier data. To address this, our study systematically analyzes the weight sensitivity of 15 MCDM methods, focusing on their application in multidisciplinary design optimization (MDO) for two automotive structures. We first introduce the concept of critical weight and investigate its underlying mechanisms, revealing that MCDM methods can be classified into two distinct categories: linear methods (e.g., WSM), which exhibit significant sensitivity to critical weights under uniform distributions or strongly conflicting objectives, and nonlinear methods (e.g., VIKOR), which demonstrate greater robustness due to their nonlinear aggregation mechanisms, etc. Based on these findings, we propose a robust decision-making framework that integrates analysis of objective distribution uniformity, conflict quantification, critical weight calculation, and robustness validation. When applied to the MDO of a battery pack system, this framework reduces the decision failure rate to 9.1%, significantly improving stability. Our study thus provides a data-driven methodology for selecting and weighting MCDM methods in automotive MDO, enhancing both reliability and practicality.