Adaptive accuracy trust region: using cross-validation in the optimization process

This article presents an adaptive accuracy trust region (AATR) optimization strategy where cross-validation is used by the trust region to reduce the number of sample points needed to construct metamodels for each step of the optimization process. Lower accuracy metamodels are initially used for the larger trust regions, and higher accuracy metamodels are used for the smaller trust regions towards the end of optimization. Various metamodelling strategies are used in the AATR algorithm: optimal and inherited Latin hypercube sampling to generate experimental designs; quasi-Newton, kriging and polynomial regression metamodels to approximate the objective function; and the leave-k-out method for validation. The algorithm is tested with two-dimensional single-discipline problems. Results show that the AATR algorithm is a promising method when compared to a traditional trust region method. Polynomial regression in conjunction with a new hybrid inherited-optimal Latin hypercube sampling performed the best.     

Design of manufacturing systems by a hybrid approach with neural network metamodelling and stochastic local search

Metamodels are models of simulation models. Metamodels are able to estimate the simulation responses corresponding to a given combination of input variables. A simulation metamodel is easier to manage and provides more insights than simulation alone. Traditionally, the multiple regression analysis is utilized to develop the metamodel from a set of simulation experiments. Simulation can consequentially benefit from the metamodelling in post-simulation analysis. A backpropagation (BP) neural network is a proven tool in providing excellent response predictions in many application areas and it outperforms regression analysis for a wide array of applications. In this paper, a BP neural network is used to generate metamodels for simulated manufacturing systems. For the purpose of optimal manufacturing systems design, mathematical models can be formulated by using the mapping functions generated from the neural network metamodels. The optimization model is then solved by a stochastic local search approach, simulated annealing (SA), to obtain an optimal configuration with respect to the objective of the systems design. Instead of triggering the detailed simulation programs, the SA-based optimization procedure evaluates the simulation outputs by the neural network metamodels. By using the SA-based optimization algorithm, the solution space of the studied problem is extensively exploited to escape the entrapment of local optima while the number of time consuming simulation runs is reduced. The proposed methodology is illustrated to be both effective and efficient in solving a manufacturing systems design problem through an example.     

Ensemble of Surrogates for Dual Response Surface Modeling in Robust Parameter Design

The robust parameter design of industrial processes and products on the basis of the concept of building quality into a design has attracted much attention from researchers and practitioners for many years, and several methods have been studied in the research community. Dual response surface methodology is one of the most commonly used approaches for simultaneously optimizing the mean and the variance of response in quality engineering. Nevertheless, when the relationship between influential input factors and output quality characteristics of a process is very complex (e.g. highly nonlinear and noisy), traditional approaches have their limitations. In this article, we introduced support vector regression, kriging model, and radial basis function, which are commonly used in computer experiments, into robust parameter design, and especially introduced a new strategy that builds the dual response surface using the ensemble of surrogates, which can provide a more robust approximation model. We demonstrated the advantages of kriging, support vector regression, radial basis function, and the ensemble of surrogates by reinvestigating the dual response approach on the basis of parametric, nonparametric, and semiparametric approaches, and a simulation experiment is studied. The results show that our presented models can achieve more desirable results than parametric, nonparametric, and semiparametric approaches in terms of fitting and predictive accuracy, and the optimal operating conditions recommended by our presented models are similar to those recommended in literature, which indicates the validation of our presented models. Copyright © 2012 John Wiley & Sons, Ltd.     

Multiobjective crashworthiness optimization design of functionally graded foam-filled tapered tube based on dynamic ensemble metamodel

Foam-filled thin-walled structures have recently gained attention with increasing interest due to their excellent energy absorption capacity. In this study, a new type of foam-filled thin-walled structure called as functionally graded foam-filled tapered tube (FGFTT) is proposed. FGFTT consists of graded density foam and thin-walled tapered tube. In order to investigate the energy absorption characteristics of FGFTTs, the numerical simulations for two kinds of FGFTTs subjected to axial dynamical loading are carried out by nonlinear finite element code LS-DYNA. In addition, a new kind of multiobjective crashworthiness optimization method employing the dynamic ensemble metamodeling method together with the multiobjective particle swarm optimization (MOPSO) algorithm is presented. This new kind of multiobjective crashworthiness optimization method is then used to implement the crashworthiness optimization design of FGFTTs. Meanwhile, the crashworthiness optimization designs of FGFTTs are implemented by using traditional multiobjective crashworthiness optimization method, which employs metamodels such as polynomial response surface (PRS), radial basis function (RBF), kriging (KRG), support vector regression (SVR) or the ensemble with the static design of experiment (DOE). Finally, by comparing the optimal designs of FGFTTs obtained by using the new multiobjective crashworthiness optimization method and the traditional one, the results show that the proposed new crashworthiness optimization method is more feasible.     

Multiple-optima search method based on a metamodel and mathematical morphology

This article investigates a non-population-based optimization method using mathematical morphology and the radial basis function (RBF) for multimodal computationally intensive functions. To obtain several feasible solutions, mathematical morphology is employed to search promising regions. Sequential quadratic programming is used to exploit the possible areas to determine the exact positions of the potential optima. To relieve the computational burden, metamodelling techniques are employed. The RBF metamodel in different iterations varies considerably so that the positions of potential optima are moving during optimization. To find the pair of correlative potential optima between the latest two iterations, a tolerance is presented. Furthermore, to ensure that all the output minima are the global or local optima, an optimality judgement criterion is introduced.     

Exploration of Metamodeling Sampling Criteria for Constrained Global Optimization

The use of surrogate models or metamodeling has lead to new areas of research in simulation-based design optimization. Metamodeling approaches have advantages over traditional techniques when dealing with the noisy responses and/or high computational cost characteristic of many computer simulations. This paper focuses on a particular algorithm, Efficient Global Optimization (EGO) that uses kriging metamodels. Several infill sampling criteria are reviewed, namely criteria for selecting design points at which the true functions are evaluated. The infill sampling criterion has a strong influence on how efficiently and accurately EGO locates the optimum. Variance-reducing criteria substantially reduce the RMS error of the resulting metamodels, while other criteria influence how locally or globally EGO searches. Criteria that place more emphasis on global searching require more iterations to locate optima and do so less accurately than criteria emphasizing local search.     

Probability-based least square support vector regression metamodeling technique for crashworthiness optimization problems

This paper presents a crashworthiness design optimization method based on a metamodeling technique. The crashworthiness optimization is a highly nonlinear and large scale problem, which is composed various nonlinearities, such as geometry, material and contact and needs a large number expensive evaluations. In order to obtain a robust approximation efficiently, a probability-based least square support vector regression is suggested to construct metamodels by considering structure risk minimization. Further, to save the computational cost, an intelligent sampling strategy is applied to generate sample points at the stage of design of experiment (DOE). In this paper, a cylinder, a full vehicle frontal collision is involved. The results demonstrate that the proposed metamodel-based optimization is efficient and effective in solving crashworthiness, design optimization problems.     

Bayesian Local Kriging

We consider the problem of constructing metamodels for computationally expensive simulation codes; that is, we construct interpolators/predictors of functions values (responses) from a finite collection of evaluations (observations). We use Gaussian process (GP) modeling and kriging, and combine a Bayesian approach, based on a finite set GP models, with the use of localized covariances indexed by the point where the prediction is made. Our approach is not based on postulating a generative model for the unknown function, but by letting the covariance functions depend on the prediction site, it provides enough flexibility to accommodate arbitrary nonstationary observations. Contrary to kriging prediction with plug-in parameter estimates, the resulting Bayesian predictor is constructed explicitly, without requiring any numerical optimization, and locally adjusts the weights given to the different models according to the data variability in each neighborhood. The predictor inherits the smoothness properties of the covariance functions that are used and its superiority over plug-in kriging, sometimes also called empirical-best-linear-unbiased predictor, is illustrated on various examples, including the reconstruction of an oceanographic field over a large region from a small number of observations. Supplementary materials for this article are available online.     

Use of Design Sensitivity Information in Response Surface and Kriging Metamodels

Metamodels based on responses from designed (numerical) experiments may form efficient approximations to functions in structural analysis. They can improve the efficiency of Engineering Optimization substantially by uncoupling computationally expensive analysis models and (iterative) optimization procedures. In this paper we focus on two strategies for building metamodels, namely Response Surface Methods (RSM) and kriging. We discuss key-concepts for both approaches, present strategies for model training and indicate ways to enhance these metamodeling approaches by including design sensitivity data. The latter may be advantageous in situations where information on design sensitivities is readily available, as is the case with e.g. Finite Element Models. Furthermore, we illustrate the use of RSM and kriging in a numerical model study and conclude with some remarks on their practical value.     

Optimizing Latin hypercube design for sequential sampling of computer experiments

Space-filling and projective properties are desired features in the design of computer experiments to create global metamodels to replace expensive computer simulations in engineering design. The goal in this article is to develop an efficient and effective sequential Quasi-LHD (Latin Hypercube design) sampling method to maintain and balance the two aforementioned properties. The sequential sampling is formulated as an optimization problem, with the objective being the Maximin Distance, a space-filling criterion, and the constraints based on a set of pre-specified minimum one-dimensional distances to achieve the approximate one-dimensional projective property. Through comparative studies on sampling property and metamodel accuracy, the new approach is shown to outperform other sequential sampling methods for global metamodelling and is comparable to the one-stage sampling method while providing more flexibility in a sequential metamodelling procedure.     

Methodology for metamodelling of microstructure evolution: precipitation kinetic case study

Numerical modeling is an important tool assisting in the designing and optimization of the production technology. The highest predictive capabilities are offered by multiscale modeling. The most important limitation of its wide application is computational cost. One of possible solutions is application of metamodels for fine scale modeling. In this paper, a systematic approach to development of metamodels is presented. All necessary steps, analyzing the model, selecting the metamodel inputs and outputs, gathering the training and testing datasets, choosing a metamodelling technique, training and testing the metamodel are described with a scientific background and practical examples. Development of the exemplary metamodel, replacing thermodynamic modeling of precipitation kinetic is presented.     

A non‐stationary covariance‐based Kriging method for metamodelling in engineering design

Metamodels are widely used to facilitate the analysis and optimization of engineering systems that involve computationally expensive simulations. Kriging is a metamodelling technique that is well known for its ability to build surrogate models of responses with non‐linear behaviour. However, the assumption of a stationary covariance structure underlying Kriging does not hold in situations where the level of smoothness of a response varies significantly. Although non‐stationary Gaussian process models have been studied for years in statistics and geostatistics communities, this has largely been for physical experimental data in relatively low dimensions. In this paper, the non‐stationary covariance structure is incorporated into Kriging modelling for computer simulations. To represent the non‐stationary covariance structure, we adopt a non‐linear mapping approach based on parameterized density functions. To avoid over‐parameterizing for the high dimension problems typical of engineering design, we propose a modified version of the non‐linear map approach, with a sparser, yet flexible, parameterization. The effectiveness of the proposed method is demonstrated through both mathematical and engineering examples. The robustness of the method is verified by testing multiple functions under various sampling settings. We also demonstrate that our method is effective in quantifying prediction uncertainty associated with the use of metamodels. Copyright © 2006 John Wiley & Sons, Ltd.     

Global response approximation with radial basis functions

In this article, a study is performed on the accuracy of radial basis functions (RBFs) in creating global metamodels for both low- and high-order nonlinear responses. The response surface methodology (RSM), which typically uses linear or quadratic polynomials, is inappropriate for creating global models for highly nonlinear responses. The RBF, on the other hand, has been shown to be accurate for highly nonlinear responses when the sample size is large. However, for most complex engineering applications only limited numbers of samples can be afforded; it is desirable to know whether the RBF is appropriate in this situation, especially when the augmented RBF has to be used. Because the types of true responses are typically unknown a priori, it is essential for high-fidelity metamodeling to have an RBF or RBFs that are appropriate for linear, quadratic, and higher-order responses. To this end, this study compares a variety of existing basis functions in both non-augmented and augmented forms with various types of responses and limited numbers of samples. This article shows that the augmented RBF models created by Wu’s compactly supported functions are the most accurate for the various test functions used in this study.     

An adaptive sampling method for variable-fidelity surrogate models using improved hierarchical kriging

Variable-fidelity (VF) modelling methods have been widely used in complex engineering system design to mitigate the computational burden. Building a VF model generally includes two parts: design of experiments and metamodel construction. In this article, an adaptive sampling method based on improved hierarchical kriging (ASM-IHK) is proposed to refine the improved VF model. First, an improved hierarchical kriging model is developed as the metamodel, in which the low-fidelity model is varied through a polynomial response surface function to capture the characteristics of a high-fidelity model. Secondly, to reduce local approximation errors, an active learning strategy based on a sequential sampling method is introduced to make full use of the already required information on the current sampling points and to guide the sampling process of the high-fidelity model. Finally, two numerical examples and the modelling of the aerodynamic coefficient for an aircraft are provided to demonstrate the approximation capability of the proposed approach, as well as three other metamodelling methods and two sequential sampling methods. The results show that ASM-IHK provides a more accurate metamodel at the same simulation cost, which is very important in metamodel-based engineering design problems.     

Considerations of accuracy and uncertainty with kriging surrogate models in single-objective electromagnetic design optimisation

The high computational cost of evaluating objective functions in electromagnetic optimum design problems necessitates the use of cost-effective techniques. The paper discusses the use of one popular technique, surrogate modelling, with emphasis placed on the importance of considering both the accuracy of, and uncertainty in, the surrogate model. After a brief review of how such considerations have been made in the single-objective optimisation of electromagnetic devices, their use with kriging surrogate models is investigated. Traditionally, space-filling experimental designs are used to construct the initial kriging model, with the aim of maximising the accuracy of the initial surrogate model, from which the optimisation search will start. Utility functions, which balance the predictions made by this model with its uncertainty, are often used to select the next point to be evaluated. In this paper, the performances of several different utility functions are examined, with experimental designs that yield initial kriging models of varying degrees of accuracy. It is found that no advantage is necessarily achieved through a search for optima using utility functions on initial kriging models of higher accuracy, and that a reduction in the total number of objective function evaluations can be achieved if the iterative optimisation search is started earlier with utility functions on kriging models of lower accuracy. The implications for electromagnetic optimum design are discussed     

Reliability-based design optimization of a car body using dimension- reduced Chebyshev polynomial

A reliability-based design optimization (RBDO) method of a car body is presented on basis of dimension-reduced Chebyshev polynomial method (DCM). To improve calculation efficiency and save computational time, complex models are often approximated by metamodels in reliability analysis. Traditional metamodels require a large number of sample points, which is time-consuming. To improve the efficiency, DCM is proposed to approximate the performance function of the car body. First, the performance function is decomposed by the dimension-reduction method into a sum of univariate functions, which are then fitted through Chebyshev polynomials. The reliability of the car body is predicted by the Taylor expansion method and the fourth-moment method. Finally, the result of RBDO is obtained using an improved adaptive genetic algorithm. The proposed method saves on the calculation time with high precision. Besides, the improved adaptive genetic algorithm reduces the number of iterations in the car body optimization and improves the efficiency.     

Non-stationary kriging for design optimization

Traditional surrogate modelling techniques, such as kriging, have been employed quite effectively within design optimizations. However, such models can fail to reproduce non-stationary responses accurately. This article explores the application of non-stationary kriging to design optimization and attempts to determine its applicability with regard to the optimization of both stationary and non-stationary objective functions. A series of analytical test problems and an engineering design problem are used to compare the performance of non-stationary and adaptive partial non-stationary kriging to traditional stationary kriging.     

一种薄壁吸能结构的设计优化

研究了一种薄壁结构在耐撞指标下的轴向冲击吸能特性。采用非线性有限元软件ABAQUS对结构的冲击过程进行仿真,并结合径向基函数法(RBF),根据耐撞性指标优化了这种吸能结构的截面,得到了较为理想的结构形式。数值结果表明采用该方法可以精确地确定优化参数,使结构比吸能(Es)得到明显提高。     

Time‐domain vector potential technique for the meshless radial point interpolation method

A time‐domain meshless algorithm based on vector potentials is introduced for the analysis of transient electromagnetic fields. The proposed numerical algorithm is a modification of the radial point interpolation method, where radial basis functions are used for local interpolation of the vector potentials and their derivatives. In the proposed implementation, solving the second‐order vector potential wave equation intrinsically enforces the divergence‐free property of the electric and magnetic fields. Furthermore, the computational effort associated with the generation of a dual node distribution (as required for solving the first‐order Maxwell's equations) is avoided. The proposed method is validated with several examples of 2D waveguides and filters, and the convergence is empirically demonstrated in terms of node density or size of local support domains. It is further shown that inhomogeneous node distributions can provide increased convergence rates, that is, the same accuracy with smaller number of nodes compared with a solution for homogeneous node distribution. A comparison of the magnetic vector potential technique with conventional radial point interpolation method is performed, highlighting the superiority of the divergence‐free formulation. Copyright © 2015 John Wiley & Sons, Ltd.     

An adaptive metamodel-based global optimization algorithm for black-box type problems

In this article, an adaptive metamodel-based global optimization (AMGO) algorithm is presented to solve unconstrained black-box problems. In the AMGO algorithm, a type of hybrid model composed of kriging and augmented radial basis function (RBF) is used as the surrogate model. The weight factors of hybrid model are adaptively selected in the optimization process. To balance the local and global search, a sub-optimization problem is constructed during each iteration to determine the new iterative points. As numerical experiments, six standard two-dimensional test functions are selected to show the distributions of iterative points. The AMGO algorithm is also tested on seven well-known benchmark optimization problems and contrasted with three representative metamodel-based optimization methods: efficient global optimization (EGO), GutmannRBF and hybrid and adaptive metamodel (HAM). The test results demonstrate the efficiency and robustness of the proposed method. The AMGO algorithm is finally applied to the structural design of the import and export chamber of a cycloid gear pump, achieving satisfactory results.