Abstract
This paper explores a new metamodeling framework that may collapse the computational explosion that characterizes the modeling of complex systems under a multiobjective and/or multidisciplinary setting. Under the new framework, a pseudo response surface is constructed for each design objective for each discipline. This pseudo response surface has the unique property of being highly accurate in Pareto optimal regions, while it is intentionally allowed to be inaccurate in other regions. In short, the response surface for each design objective is accurate only where it matters. Because the pseudo response surface is allowed to be inaccurate in other regions of the design space, the computational cost of constructing it is dramatically reduced. An important distinguishing feature of the new framework is that the response surfaces for all the design objectives are constructed simultaneously in a mutually dependent fashion, in a way that identifies Pareto regions for the multiobjective problem. The new framework supports the puzzling notion that it is possible to obtain more accuracy and radically more design space exploration capability, while actually reducing the computation effort. This counterintuitive metamodeling paradigm shift holds the potential for identifying highly competitive products and systems that are well beyond today’s state of the art.
Similar content being viewed by others
Abbreviations
- PRS:
-
Pseudo Response Surface
- RS:
-
Response Surface
- \(\tilde{(\textrm{ })}\) :
-
Denotes approximate quantity
- g :
-
Vector of inequality constraints (see (2))
- h :
-
Vector of equality constraints (see (3))
- n :
-
Number of design objectives (see (1))
- n p :
-
Number of explored nearly Pareto points (see (20))
- n f :
-
Number of pseudo points
- p k :
-
k-th utopia plane point (see (5))
- x :
-
Vector of design variables (see (1))
- x i(f) :
-
i-th pseudo design point (Fig. 8)
- \(\tilde{x}^{i(l)}\) :
-
i-th data point used to build the local approximations \(\tilde{\mu}^{(l)}(x)\) (see (14))
- x (p) :
-
Set of nearly Pareto designs (see (20))
- x pk :
-
Nearly Pareto design corresponding to the utopia plane point p k (see (20))
- μ j (x):
-
j-th computationally expensive design objective (see (1))
- \(\tilde{\mu}_{j}^{(l)}(x)\) :
-
Local approximation of the j-th design objective (see (15))
- μ (p) :
-
Set of nearly Pareto points in objective space (see (21))
- μ pk :
-
Set of nearly Pareto point in the objective space, corresponding to the utopia plane point p k (see (21))
- \(\tilde{\mu_{j}}^{\mathrm{PRS}}(x)\) :
-
Pseudo response surface model for the j-th design objective (Fig. 9)
References
Alexandrov NM, Lewis RM (2001) An overview of first-order model management for engineering optimization. Optim Eng 2:413–430
Balabanov VO, Giunta AA, Golovidov O, Grossman B, Mason WH, Watson LT, Haftka RT (1999) Reasonable design space approach to response surface approximation. J Aircr 36(1):308–315
Bokose FL, Vandevelde L, Melkebeek JAA (2004) Sequential approximate multiobjective optimisation of switched reluctance motor design using surrogate models and nongradient local search algorithm. IEE Proc Sci Meas Technol 151(6):471–475
Box GE, Draper NR (1987) Empirical model-building and response surfaces. Wiley, New York
Deb K, Pratap A, Agarwal S, Meyarivan T (2002) A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Trans Evol Comput 6(2):182–197
Giunta AA, Watson LT (1998) A comparison of approximation modeling techniques: Polynomial versus interpolating models. Technical Report AIAA-98-4758
Giunta AA, Balabanov V, Haim D, Grossman B, Mason WH, Watson LT, Haftka RT (1997) Multidisciplinary optimisation of a supersonic transport using design of experiments theory and response surface modeling. Aeronaut J 101(1008):347–356
Hosder S, Watson LT, Grossman B, Mason WH, Kim H, Haftka RT, Cox SE (2001) Polynomial response surface approximations for the multidisciplinary design optimization of high speed civil transport. Optim Eng 2(4):431–452
Hussain MF, Barton RR, Joshi SB (2002) Metamodeling: Radial basis functions, versus polynomials. Eur J Oper Res 138(1):142–154
Jin R, Chen W, Simpson T (2001) Comparative studies of metamodelling techniques under multiple modelling criteria. Struct Multidiscip Optim 23(1):1–13
Koch PN, Simpson TW, Allen JK, Mistree F (1999) Statistical approximations for multidisciplinary design optimization: the problem of size. J Aircr 36(1):275–286
Leary SJ, Bhaskar A, Keane AJ (2001) A constraint mapping approach to the structural optimization of an expensive model using surrogates. Optim Eng 2:385–398
Mattson CA, Mullur AA, Messac A (2004) Smart Pareto filter: obtaining a minimal representation of multiobjective design space. Eng Optim 36(6):721–740
Messac A, Mattson CA (2004) Normal constraint method with guarantee of even representation of complete Pareto frontier. AIAA J 42(10):2101–2111
Messac A, Ismail-Yahaya A, Mattson CA (2003) The normalized normal constraint method for generating the Pareto frontier. Struct Multidiscip Optim 25(2):86–98
Messac A, Van Dessel S, Mullur AA, Maria A (2004) Optimization of large scale rigidified inflatable structures for housing using physical programming. Struct Multidiscip Optim 26(1–2):139–151
Mullur AA, Messac A (2005a) Extended radial basis functions: more flexible and effective metamodeling. AIAA J 43(6):1306–1315
Mullur AA, Messac A (2005b) Higher metamodel accuracy using computationally efficient pseudo response surfaces. In: 46th AIAA/ASME/ASCE/AHS/ASC structures, structural dynamics, and materials conference AIAA-2005-2058, Austin, TX, April 2005
Mullur AA, Messac A (2005c) Multiobjective structural optimization using computationally efficient pseudo response surfaces. In: 6th world congress on structural and multidisciplinary optimization, Rio de Janeiro, May–June 2005
Myers RH, Montgomery DC (2002) Response surface methodology, 2nd edn. Wiley–Interscience, New York
Nakayama H, Arakawa M, Sasaki R (2002) Simulation-based optimization using computational intelligence. Optim Eng 3(2):201–214
Papadrakakis M, Lagaros ND, Tsompanakis Y (1998) Structural optimization using evolution strategies and neural networks. Comput Methods App Mech Eng 156(1–4):309–333
Renaud JE, Gabriele GA (1993) Improved coordination in non-hierarchic system optimization. AIAA J 31(12):2367–2373
Rodriguez JF, Renaud JE, Wujek BA, Tappeta RV (2000) Trust region model management in multidisciplinary design optimization. J Comput Appl Math 124:139–154
Rodriguez JF, Perez VM, Padmanabhan D, Renaud JE (2001) Sequential approximate optimization using variable fidelity response surface approximations. Struct Multidiscip Optim 22:24–34
Sakata S, Ashida F, Zako M (2003) Structural optimization using kriging approximation. Comput Methods Appl Mech Eng 192:923–939
Simpson TW, Peplinski JD, Koch PN, Allen JK (2001) Metamodels for computer-based engineering design: Survey and recommendations. Eng Comput 17:129–150
Sobester A, Leary SJ, Keane AJ (2004) A parallel updating scheme for approximating and optimizing high fidelity computer simulations. Struct Multidiscip Optim 27:371–383
Sobieski IP, Kroo IM (2000) Collaborative optimization using response surface estimation. AIAA J 38(10):1931–1938
Vanderplaats RD (2001) Genesis: structural analysis and optimization, user manual, vol II, August 2001. Version 7.0
Van Veldhuizen DA, Lamont GB (1999) Multiobjective evolutionary algorithm test suites. In: Proceedings of the ACM symposium on applied computing, San Antonio, TX, February–March 1999, pp 351–357
Wilson B, Cappelleri D, Simpson WT, Frecker M (2001) Efficient Pareto frontier exploration using surrogate approximations. Optim Eng 2:31–50
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Messac, A., Mullur, A.A. A computationally efficient metamodeling approach for expensive multiobjective optimization. Optim Eng 9, 37–67 (2008). https://doi.org/10.1007/s11081-007-9008-0
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11081-007-9008-0