摘要
为改善铸件生产质量,增强参数设计稳健性,提出一种基于双向投影决策的参数优化方法。首先,将犹豫模糊集引入参数优化过程中,构建犹豫模糊决策矩阵,更确切描述产品在噪声干扰下形成的质量波动,使之更加贴近实际生产过程。之后,通过考虑各组工艺方案与其正、负理想点之间的接近程度,计算Hausdorff距离,将多个响应转化为双投影后的贴近度,使多响应系统得以简化,提高系统简化结果的可靠性。然后将主效应分析与神经网络相结合进行参数设计,获得最优参数组合。最后将所研究方法应用于铸造生产工艺中,改善铸件生产质量,验证了本文所提方法的有效性。
In order to improve the quality of casting production and enhance its robustness,a parameter optimization approach based on bidirectional projection decision was proposed.Firstly,the hesitant fuzzy set was introduced into the parameter optimization process to construct the hesitant fuzzy decision matrix,which could more accurately describe the quality fluctuations in the form of noise interference,and make it closer to the real production process.Next,considering the proximity between each project and its positive or negative ideal point,the Hausdorff distance was calculated.It transformed multiple responses into the proximity degree through bidirectional projection,which simplified the multi-response system and improved the reliability of result correspondingly.Furthermore,the main effect analysis and neural network were combined to obtain the optimal combination of parameters.Finally,a case study of quality improvement for the casting production was provided,which demonstrated the effectiveness of the proposed approach.
作者
崔庆安
谷丰盈
禹建丽
CUI Qing′an;GU Fengying;YU Jianli(School of Economics and Management,Shanghai Maritime University,Shanghai 201306,China;School of Management Engineering,Zhengzhou University of Aeronautics,Zhengzhou,Henan 450046,China)
出处
《工业工程与管理》
北大核心
2022年第1期157-163,共7页
Industrial Engineering and Management
基金
国家自然科学基金面上项目(71571168)
国家自然科学基金河南省联合基金重点项目(U1904211)
国家创新方法工作专项(2019IM020200)
河南省高校科技创新人才支持计划(2019cx007)
河南省科技攻关计划项目(182102210107)
关键词
双向投影决策
多响应参数优化
神经网络
铸造工艺
bidirectional projection decision
multi-response parameter optimization
neural network
casting process
作者简介
崔庆安(1974—),山西襄垣人,教授,博士,主要研究方向为质量管理与质量优化、复杂系统建模及优化。E-mail:qacui@shmtu.edu.cn。
