摘要
针对产品设计中普遍存在的不确定性问题,提出一种贝叶斯与满意度函数相结合的处理方法.首先,采用贝叶斯理论分析各参数并根据后验分布进行随机抽样,依据工程人员偏好的满意度大小,优化满意度函数以获得满意区间,从而保证所求优化解在不确定下仍以较高的满意度散布在可行域内;其次,以目标在规格限内的后验概率为优化模型,在满意区间内最大化优化模型以得到优化解;最后,结合实际工业案例验证所提方法能够有效地处理产品设计问题中不确定性源对结果的影响.
As to uncertainty which widely exists in product design problems,this paper proposed a method combining Bayesian theory and desirability function. Firstly,Bayesian theory was used to analyze parameters and sampled them according to posterior distribution and then the desirability function was optimized based on the desirability of engineers' preference size in order to obtain satisfied interval,which could guarantee that the optimal variable settings have high desirability in feasible region under parameter uncertainty. Secondly,a criterion that is the probability that all objectives simultaneously meet their respective specifications was constructed and that criterion was optimized in satisfied feasible region to obtain the optimal variable settings. Finally,a practical industrial example was used to verify the effectiveness of the proposed method to solve the robust optimization design under parameter uncertainty.
出处
《安徽师范大学学报（自然科学版）》
CAS
2015年第4期337-342,共6页
Journal of Anhui Normal University(Natural Science)
基金
安徽高校省级自然科学基金(KJ2009B014Z)
关键词
贝叶斯分析
不确定性
满意度函数
Bayesian analysis
uncertainty
desirability function
