基于语义的产品自适应混合免疫遗传形态设计

宋红; 余隋怀; 耶琳; 赵明磊

doi:10.13433/j.cnki.1003-8728.2014.1104

基于语义的产品自适应混合免疫遗传形态设计

doi: 10.13433/j.cnki.1003-8728.2014.1104

宋红¹,
余隋怀^1, ,,
耶琳²,
赵明磊¹

1.
西北工业大学工业设计研究所, 西安 710072;
2.
摩托罗拉系统有限公司创新设计部, 佛罗里达州 33322

基金项目:

国家自然科学基金项目(51105310)资助

详细信息

作者简介:
宋红(1977- ),博士研究生,研究方向为数字媒体艺术设计和计算机辅助工业设计(CA1D),hsong@xsyu.edu.cn。

通讯作者:
余隋怀,教授,博士生导师,ysuihuai@vip.sina.com.

计量
- 文章访问数: 125
- HTML全文浏览量: 29
- PDF下载量: 2
- 被引次数: 0
出版历程
- 收稿日期: 2013-03-11

A Semantic-based Adaptive Hybrid Immune Product Shape Design Method

1.
Institute of Industrial Design, Northwestern Polytechnical University, Xi'an 710072;
2.
Innovation Design, Motorola Solutions Inc., Plantation, FL USA 33322

摘要

摘要: 为进一步探讨产品设计中形态特征和用户情感匹配的问题,提出了一种基于语义的自适应混合免疫遗传产品形态设计优化方法.通过目标语义的模糊隶属度的匹配来生成初始抗体集合,再利用专家建议来生成疫苗集合,并在演化过程中通过疫苗接种和基于亲和度的选择来加速收敛并保持种群多样性,利用反向传播神经网络(BPNN)将形态特征语义空间映射到情感语义空间并构造适应度函数.此外,疫苗接种概率按接种效果进行自适应调节,并基于信息熵理论定义抗体之间的亲合度及抗体的选择概率.原型系统实现了形态设计的快速生成和设计评价,验证了方法的有效性.
- 免疫 /
- 遗传算法 /
- 反向传播神经网络 /
- 形态设计
Abstract: To investigate the shape characteristic and the user emotion in the product shape design process,a semantic-based adaptive hybrid immune genetic design approach is proposed. The initial antibody set was generated as a result of matching with the target design fussy definition. Then the extracted vaccine set was created based on the experts'feedback and evaluation. During the evolution process,the antibodies convergence was accelerated and the diversity was maintained by the vaccination and affinity-based selection. Also,the fitness function was defined with a back-propagation neural network which maps the semantic space of the shape characteristic design to the semantic space of the emotion. In addition,the vaccination possibilities of vaccines adapt to the effects of vaccination and the affinities and selection possibilities of antibodies were defined according to the information entropy theory. The results of prototype system demonstrated that the present method can efficiently generate and evaluate product shape design scheme.
- antibodies /
- back propagation neural network /
- backpropagation algorithms /
- design

HTML全文

参考文献(11)

[1]	Sevaux M,Mineur Y. A curve-fitting genetic: algorithmfor a styling application[J]European Journal of Operational Research,2007,179(3):895-905
[2]	Li Y F,Zhu L P. Research on producta image formoptimization design[C]//Computer Seience andInforniation Technology. IEEE International Conference onComputer Seience and Inforniation Technology, 2010;9-11,2010,IEEE,2010}173-177
[3]	Hsiao SW,Chiua F Y, Lua S H. Producta-form designmodel based on genetic: algorithms[J]InternationalJournal of Industrial Ergonomic、,2010,40(3);237-246
[4]	Heermann P D, Khazenie N. Classification of multispectaralremote sensing data using a back-propagation neuralnetwork[J]IEEE Trans. on Geoseience and RemoteSensing,1992,30(1):81-88
[5]	Goldberg D E. Genetic: algorithms in search, optimization,and machine learning[M]New Jersey:Addison-Wesley,1989
[6]	Song H,Zhang D,Liu J. An adaptive hybrid immunegenetic: algorithm for maximum cut problem[J]. LectaureNotes in Computer Seience,2005:863-866
[7]	Jiao L C,Wang L. A novel genetic: algorithm based onimmunity[J]IEEE Trans. on Systems,Man andCybernetic、,2000,30(5);552-561
[8]	Chun J S, Kim M K, Jung H K, et al. Shape optinuzationof electaromagnetic devices using immune algorithm[J]IEEE Trans. on Magnetic,1997,33(2):1976-1879
[9]	Srinivas M, Patnaik L M. Adaptive probabilities of crossoverand mutations in gas[J]. IEEE Trans. on Systems,Manand Cybernetic、,1994,24(4);493-530
[10]	徐江,孙守迁基于正交-交互式遗传算法的产品形态设计[J]计算机集成制造系统,2007,13(8);1470-1475Xu J,Sun S Q. Producta form design based onorthogonal interactaive genetic: algorithm[J]ComputerIntegrated Manufactauring Systems,2007,13(8):1470-1475(in Chinese)
[11]	Zwicky F, Invention D. Research through the morphologicalapproach[J]New Methods of Thought and Procedure,1967:273-297"