Metamodel-assisted optimization based on multiple kernel regression for mixed variables

详细信息    查看全文

• 作者：Manuel Herrera (1)
  Aurore Guglielmetti (2)
  Manyu Xiao (2)
  Rajan Filomeno Coelho (1)
  
• 关键词：Multiple kernel regression ; Mixed variables ; Metamodels ; Categorical variables ; Dummy coding
• 刊名：Structural and Multidisciplinary Optimization
• 出版年：2014
• 出版时间：June 2014
• 年：2014
• 卷：49
• 期：6
• 页码：979-991
• 全文大小：
• 参考文献：1. Abramson M, Audet C, Dennis DEJ (2004) Filter pattern search algorithms for mixed variable constrained optimization problems. SIAM J Optim 11:573鈥?94
  2. Agresti A (1996) An introduction to categorical data analysis. Wiley, New York
  3. Christmann A, Hable R (2012) Consistency of support vector machines using additive kernels for additive models. Comput Stat Data Anal 56(4):854鈥?73 CrossRef
  4. Coello Coello CA, Van Veldhuizen DA, Lamont GB (2002) Evolutionary algorithms for solving multi-objective problems. Kluwer Academic/Plenum Publishers, New York CrossRef
  5. Cohen J, Cohen P, West SG, Aiken LS (2003) Applied multiple regression/correlation analysis for the behavioural sciences. Routledge, New York
  6. Davis MJ (2010) Contrast coding in multiple regression analysis: strengths, weaknesses, and utility of popular coding structures. Data Sci 8:61鈥?3
  7. Deb K, Pratap A, Agarwal S, Meyarivan T (2002) A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Trans Evol Comput 6(2):182鈥?97 CrossRef
  8. Ferreira AJM (2009) MATLAB codes for finite element analysis. Solid mechanics and its applications. Springer, New York
  9. Filomeno Coelho R (2012) Extending moving least squares to mixed variables for metamodel-assisted optimization. In: 6th European congress on computational methods in applied sciences and engineering (ECCOMAS 2012). Vienna
  10. Filomeno Coelho R (2013) Metamodels for mixed variables based on moving least squares鈥揳pplication to the structural analysis of a rigid frame. Optim Eng. doi:10.1007/s11081-013-9216-8
  11. Forrester AIJ, Keane AJ (2009) Recent advances in surrogate-based optimization. Prog Aerosp Sci 45(1鈥?):50鈥?9 CrossRef
  12. Goldberg Y, Elhadad M (2008) splitSVM: Fast, space-efficient, non-heuristic, polynomial kernel computation for NLP applications. In: 46st annual meeting of the association of computational linguistics (ACL)
  13. G枚nen M, Alpaydin E (2011) Multiple kernel learning algorithms. Mach Learn Res 12:2211鈥?268
  14. Hardy M (1993) Regression with dummy variables. Sage, Newbury Park
  15. Hemker T (2008) Derivative free surrogate optimization for mixed-integer nonlinear black box problems in engineering. PhD thesis, Technisen Universit盲t Darmstad, Germany
  16. Herrera M, Filomeno Coelho R (2013) Metamodels for mixed variables by multiple kernel regression. In: 10th world congress on structural and multidisciplinary optimization (WCSMO 10). Orlando
  17. Hofmann T, Sch枚lkopf B, Smola A (2008) Kernel methods in machine learning. Ann Stat 36(3):1171鈥?220 CrossRef
  18. Huang CM, Lee YJ, Lin DK, Huang SY (2007) Model selection for support vector machines via uniform design. Comput Stat Data Anal 52(1):335鈥?46 CrossRef
  19. Kondor RI, Lafferty JD (2002) Diffusion kernels on graphs and other discrete input spaces. In: Proceedings of the nineteenth international conference on machine learning, ICML 鈥?2. Morgan Kaufmann Publishers Inc., San Francisco, pp 315鈥?22
  20. Lanckriet G, Cristianini N, Barlett P, El-Ghaoui L, Jordan MI (2004) Learning the kernel matrix with semi-definite programming. Mach Learn Res 5:27鈥?2
  21. Lee N, Kim JM (2010) Conversion of categorical variables into numerical variables via bayesian network classifiers for binary classifications. Comput Stat Data Anal 54(5):1247鈥?265 CrossRef
  22. Liew R, Chen H, Shanmugam N, Chen W (2000) Improved non-linear plastic hinge analysis of space frame structures. Eng Struct 22(10):1324鈥?338 CrossRef
  23. Luts J, Molenberghs G, Verbeke G, Huffel SV, Suykens JA (2012) A mixed effects least squares support vector machine model for classification of longitudinal data. Comput Stat Data Anal 56(3):611鈥?28 CrossRef
  24. McCane B, Albert MH (2008) Distance functions for categorical and mixed variables. Pattern Recogn Lett 29(7):986鈥?93 CrossRef
  25. Mortier F, Robin S, Lassalvy S, Baril C, Bar-Hen A (2006) Prediction of Euclidean distances with discrete and continuous outcomes. Multivar Anal 97(8):1799鈥?814 CrossRef
  26. Papadrakakis M, Lagaros N, Plevris V (2005) Design optimization of steel structures considering uncertainties. Eng Struct 27:1408鈥?418 CrossRef
  27. Purcell R (2011) Machine learning with multiple kernel learning algorithms. Master鈥檚 thesis, University of Bristol, UK
  28. Qiu S, Lane T (2005) Multiple kernel learning for support vector regression. Tech. rep. Computer Science Department, University of New Mexico, Albuquerque
  29. Queipo NV, Haftka RT, Shyy W, Goel T, Vaidyanathan R, Tucker PK (2005) Surrogate-based analysis and optimization. Prog Aerosp Sci 41:1鈥?8 CrossRef
  30. Sch枚lkopf B (2000) The kernel trick for distances. Tech. rep., Microsoft Research
  31. Sch枚lkopf B, Smola AJ (2001) Learning with kernels, support vector machines, regularization, optimization, and beyond. MIT Press, Cambridge
  32. Shawe-Taylor J, Cristianini N (2006) Kernel methods for pattern analysis. Cambridge University Press, Cambridge
  33. Smola A, Sch枚lkopf B (2004) A tutorial on support vector regression. Stat Comput 14(3):199鈥?22 CrossRef
  34. Sonnenburg S, R盲tsch G, Sch盲fer C (2006) A general and efficient multiple kernel learning algorithm. In: Weiss Y, Sch枚lkopf B, Platt J (eds) Advances in neural information processing systems 2006. MIT Press, Cambridge, pp 1273鈥?280
  35. Tsang IW, Kwok JT, Bay CW (2003) Distance metric learning with kernels. In: International conference on artificial neural networks 2003, pp 126鈥?29
  36. Wendorf CA (2004) Primer on multiple regression coding: common forms and the additional case of repeated contrasts. Underst Stat 3:47鈥?7 CrossRef
  
• 作者单位：Manuel Herrera (1)
  Aurore Guglielmetti (2)
  Manyu Xiao (2)
  Rajan Filomeno Coelho (1)
  
  1. ULB鈥揃ATir Department, Universit茅 libre de Bruxelles, Avenue F.D. Roosevelt, 50 (CP 194/2), B-1050, Brussels, Belgium
  2. NPU鈥揇epartment of Applied Mathematics, Northwestern Polytechnical University, Xi鈥檃n, Shaanxi, 710072, People鈥檚 Republic of China
  
• ISSN：1615-1488

文摘

While studies in metamodel-assisted optimization predominantly involve continuous variables, this paper explores the additional presence of categorical data, representing for instance the choice of a material or the type of connection. The common approach consisting in mapping them onto integers might lead to inconsistencies or poor approximation results. Therefore, an investigation of the best coding is necessary; however, to build accurate and flexible metamodels, a special attention should also be devoted to the treatment of the distinct nature of the variables involved. Consequently, a multiple kernel regression methodology is proposed, since it allows for selecting separate kernel functions with respect to the variable type. The validation of the advocated approach is carried out on six analytical benchmark test cases and on the structural responses of a rigid frame. In all cases, better performances are obtained by multiple kernel regression with respect to its single kernel counterpart, thereby demonstrating the potential offered by this approach, especially in combination with dummy coding. Finally, multi-objective surrogate-based optimization is performed on the rigid frame example, firstly to illustrate the benefit of dealing with mixed variables for structural design, then to show the reduction in terms of finite element simulations obtained thanks to the metamodels.