China Mechanical Engineering ›› 2015, Vol. 26 ›› Issue (20): 2763-2770.

Previous Articles     Next Articles

Multi-objective  Optimization Method of  Composite Imagery-oriented  Product  Form

Chen Guodong1,2;Chen Siyu1;Wang Jun1;Sun Shouqian2;Ni Yihua1   

  1. 1.Zhejiang A&F University,Hangzhou,311300
    2.Zhejiang University,Hangzhou,310027
  • Online:2015-10-25 Published:2015-10-20

面向复合意象的产品形态多目标优化

陈国东1,2;陈思宇1;王军1;孙守迁2;倪益华1   

  1. 1.浙江农林大学,杭州,311300
    2.浙江大学,杭州,310027
  • 基金资助:
    国家自然科学基金资助项目(61175125,51305077);浙江省自然科学基金资助项目(LZ15E050003);浙江省高等学校访问学者教师专业发展项目(FX2014034);浙江省教育厅科研项目(Y201533202)

Abstract:

To help the product development team satisfy consumer perceptions when they designed new product,a  composite imagery oriented multi-objective optimization model for product form design was proposed.Firstly,composite imagery were elicited by semantic  differential(SD) and factor analysis(FA).Secondly,components and candidate elements of product were extracted by morphological analysis(MA).Constructing  BP  neural network between imagery and form feature,the multi-objective  genetic algorithmⅡ(NSGAⅡ) was employed to search Pareto optimum finally.Results  of soymilk machine case show  that  Pareto optimum distribute uniformly,product developer  can   select  optimization solutions with multiple objective interactively for specific circumstances, and the proposed method has a good directive function and practical value in product design area.

Key words: composite imagery, product design, multiple objective , optimization;back , propagation(BP) , neural network

摘要:

为了使产品开发人员和企业开发新产品时符合用户的多维度感性需求,提出了一种面向用户复合意象的产品形态多目标优化设计方法。首先,应用语义差异评价和因子分析法抽取用户的复合意象维度;然后,采用形态分析法提取产品部件和外形单元要素;最后,在构建产品形态的BP意象预测模型的基础上,采用多目标遗传算法NSGAII求解最优方案。豆浆机实例表明,该方法适用于复合意象的产品形态优化设计,成功得到分布均匀的Pareto最优解,产品开发者可根据具体情况在多个目标之间交互选择优化方案,具有很好的指导作用和实际应用价值。

关键词: 复合意象, 产品设计, 多目标优化, BP网络

CLC Number: