[1]张锦华,原思聪,张晓钟,等.基于人工神经网络响应面技术的多学科优化方法研究及应用[J].西安建筑科技大学学报:自然科学版,2011,43(03):451-456.[doi:DOI :10.15986/j .1006-7930.2011.03.008]
　ZH ANGJ in-hua,YUAN S i-cong,Z H ANG X iao-zhong,et al.Research and application of multidisciplinary design optimization based on artificial neural network response surface[J].J.Xi’an Univ. of Arch. & Tech.:Natural Science Edition,2011,43(03):451-456.[doi:DOI :10.15986/j .1006-7930.2011.03.008]
点击复制

基于人工神经网络响应面技术的多学科优化方法研究及应用()
分享到：

西安建筑科技大学学报:自然科学版[ISSN:1006-7930/CN:61-1295/TU]

卷:
43
期数:
2011年03期
页码:
451-456
栏目:
出版日期:
2011-06-30

文章信息/Info

Title:
Research and application of multidisciplinary design optimization based on artificial neural network response surface
文章编号:
1006-7930(2011)03-0451-06
作者:
张锦华;  原思聪;  张晓钟;  郭海燕
(西安建筑科技大学机电工程学院, 陕西西安710055)
Author(s):
ZH ANGJ in-hua ; YUAN S i-cong ;  Z H ANG X iao-zhong ; GUO Hai-y an
(Schoo l o f Mechanical and Electro nic Enginee ring , Xi′an Univ .of Ar ch .& Tech ., Xi′an 710055 , China)
关键词:
人工神经网络; 响应面; 多学科优化; 齿轮减速箱
Keywords:
arti f ical neural network ; response sur f ace ; multidiscip linary design optimiz ation ; gearbox .
分类号:
TP301.6
DOI:
DOI :10.15986/j .1006-7930.2011.03.008
文献标志码:
A
摘要:
在常用的非层次型多学科综合优化算法基础上, 提出了基于神经网络响应面多学科优化算法(ANN
MDO), 是一种二级结构优化方法, 子学科在优化时只需满足本学科的局部约束, 学科层优化目标是使该学科
优化设计方案与系统层提供的目标方案差异最小, 系统层提供一种协调各个学科优化结果冲突机制, 并且所
需学科层信息通过神经网络响应面获取.最后将协同优化算法(CO)、并行子空间优化算法(CSSO)、ANN
MDO 算法应用于齿轮减速箱算例, 验证了本文算法的高效性.
Abstract:
On the ba sis o f no n-hie rarchical multidisciplinary optimiza tion algo rithm , this paper pro po sed a new multidisciplina
ry desig n optimization based on the neural netw or k re sponse surface (ANN MDO), a two-leve l optimizatio n architecture
.That is to say , the sub-discipline lev el o nly meet the lo cal constraints and the objective is to g et smalle st difference
betw een local o ptimal so lutio n and targe t pr og ram pro vided by sy stem level .Meanw hile , the system level no t only o ffe rs
some co or dinating mechanism to g uarantee ag reement of all discipline level o ptimal solution , but also o btains discipline
level information by ar tificial neural netwo rk -based re sponse surface appro xima tion .Finally , a g earbox is ado pted as an
e xample to v erify the efficiency o f ANN MDO algo rithm , which compare the collabo rativ e o ptimiza tion (CO)with co ncurrent
subspace optimizatio n (CSSO).

参考文献/References:

References
[ 1] 　GIESING Jo seph P, BART HELEM Y Jean-Francois M .A Summar y of Industry MDO Applicatio n and Needs [ C] ∥
AIAA Technical Repo r t 1998.
[ 2] 　SOBIESKI J.Optimizatio n by Decom po sitio n:A S tep f rom Hierar chic to Nonhierar chic Sy stem .Seco nd NASA/ Air
Force Sympo sium o n Recent Adva nces in Multidisciplina ry Analy sis and Optimazatio n, NASA/ TM-101494 , 1989 .
[ 3] 　杨俊芬, 顾　强, 苏明周.基于增量动力分析和人工神经网络计算结构影响系数和位移放大系数[ J] .西安建筑科
技大学学报:自然科学版, 2010, 42(6):815-822.
YANG Jun-fen, GU Qiang , SU Ming-zhou .Calculating R and Cd based o n IDA and ANN[ J] .J.Xi′an Univ .o f
Arch .& Tech .:Na tur al Science Editio n, 2010 , 42(6):815-822 .
[ 4] 　FRANCO S .Unive rsal Appro ximation Using Feed-Forwar d Neura l Netw o rks:A Sur vey of Some Ex isting Methods
and Some New Results .Neural Netw or ks, 1998 .
[ 5] 　陈柏鸿, 肖人彬, 刘继红, 等.复杂产品协同优化设计中的耦合因素的研究[ J] .机械工程学报, 2001, 37(1):19-23.
CH EN Bai-hong , XIAO Ren-bin , LIU Ji-hong , et al .Re sear ch o n the coupled facto rs in MDO for complex pro duc
ts[ J] .Chinese Jour nal of Mechanical Eng ineering , 2001 , 37(1):19-23.
[ 6] 　钟毅芳, 陈柏鸿, 等.多学科综合优化设计原理与方法[ M] .武汉:华中科技大学出版社, 2007.
ZHONG Yi-fa ng , CHEN Bai-ho ng , et al.Principle s and metho ds of multidisciplinary design optimizatio n[ M] .Wuhan:
Huazho ng Unive rsity o f Science and Technolo gy Press, 2007 .
[ 7] 　SIMPSON T W, PEPLINSKI J D, KOCH P N, et al.M etamodels for computer-based engineering de sig n :survey
and recommendatio ns [ J] .Engineering with Compute rs , 2001 .

相似文献/References:

[1]华 军,武霞霞,李东波,等.基于神经网络的石墨烯弹性参量识别方法研究[J].西安建筑科技大学学报:自然科学版,2015,47(05):760.[doi:DOI：10.15986/j.1006-7930.2015.05.026]
　HUA Jun,WU Xiaxia,LI Dongbo,et al.Idenfication of elastic parameters method for graphene based on neural network[J].J.Xi’an Univ. of Arch. & Tech.:Natural Science Edition,2015,47(03):760.[doi:DOI：10.15986/j.1006-7930.2015.05.026]

备注/Memo

备注/Memo:
＊收稿日期:2010-07-20 　　修改稿日期:2011-04-12
基金项目:“ 十二五” 国家科技支撑项目(2010BAE00372-2);陕西省教育厅自然科学专项(09 JK559)
作者简介:张锦华(1970-), 女,陕西渭南人, 博士研究生, 讲师, 主要从事智能计算, 机械设计及理论研究.
更新日期/Last Update: 2015-11-02