关键词: 模块划分, 层次分析法, 稳健设计, 人工免疫改进算法, 多目标优化

Abstract: For module division problem of product platform design,the robust product module division was proposed by introducing the robust design idea.Through the components-correlation matrix and function-contribution matrix,the criterion function and noise function in robust product platform design was constructed.The traditional artificial immune algorithm was improved with vaccine inoculation strategy,which was used to solve the problems in multi-objective optimization.Non-dominated set in multi-objective optimization problem was taken as the vaccine to inject into offspring according to the binary gene selection mechanism,which could increase the probability of obtaining good genes for offspring.The application examples of product module division in some railway carriages were given to verify the validity and superiority of proposed method by comparing to several traditional multi-objective optimization algorithms.

Key words: module division, analytic hierarchy process, robust design, improved artificial immune algorithms, multi-objective optimization