Recent advances in surrogate-based optimization

Forrester, Alexander I. J.; Keane, Andy J.

doi:10.1016/j.paerosci.2008.11.001

Cited by 1,815 publications

(971 citation statements)

References 74 publications

Supporting

Mentioning

954

Contrasting

Unclassified

Order By: Relevance

“…= lift, drag, and pitching moment coefficients O VER the past two decades, surrogate-based modeling and optimization have played an increasingly important role in different areas of aerospace engineering [1][2][3][4][5][6][7][8], such as aerodynamic design optimization [9,10], structural optimization [11], and multidisciplinary design optimization of aircraft or spacecraft [12], where high-fidelity and expensive numerical simulations, such as computational fluid dynamics (CFD) or computational solid dynamics (CSD), are often employed. Surrogate model is also called "response surface model", "metamodel", "approximation model", or "emulator" by the literature in different research areas.…”

mentioning

confidence: 99%

“…Gradient-enhanced kriging (GEK) denotes the extension of kriging to models in which inexpensive gradient information is incorporated [47]. Note that GEK was also viewed as one kind of cokriging (e.g., [44,45]), although the terminology "cokriging" generally means the kriging incorporating lower-fidelity auxiliary data (see [5] or [33]). There are two ways of incorporating gradients into the construction of a kriging, which leads to indirect GEK and direct GEK [47].…”

mentioning

confidence: 99%

“…The drawbacks of this method are that there is a chance for numerical errors to be introduced while estimating the additional function values from the gradients [44], and it is less robust because the condition number of the correlation matrix can be very large. In case of the direct GEK [5,45,47,[51][52][53], the gradients are directly included in the kriging equations system, assuming that the predictor is defined by augmenting the weighted sum of the functional values with the weighted sum of the gradients. The optimal weights are then obtained by minimizing the mean-squared error (MSE) of the predictor, and the resulting linear equations system includes extra correlations between the function values and the gradients as well as the correlations between the gradients themselves.…”

mentioning

confidence: 99%

“…Therefore, the applications of this method are usually limited to either two-dimensional aerodynamic configurations [61] or three-dimensional (3-D) configurations using low-fidelity and fast CFD methods [62]. And the third approach is surrogate-based optimization (SBO) [1][2][3][4][5][6][7][8][9]. As previously mentioned, the surrogate models enable us to find the global optimum within a very limited number of expensive evaluations [30,38].…”

mentioning

confidence: 99%

See 3 more Smart Citations

Weighted Gradient-Enhanced Kriging for High-Dimensional Surrogate Modeling and Design Optimization

et al. 2017

View full text Add to dashboard Cite

A novel formulation of gradient-enhanced surrogate model, called weighted gradient-enhanced kriging, is proposed and used in combination with the cheap gradients obtained by the adjoint method to ameliorate the "curse of dimensionality". The core idea is to build a series of submodels with much smaller correlation matrices and then sum them up with appropriate weight coefficients, aiming to avoid the prohibitive cost associated with decomposing the large correlation matrix of a gradient-enhanced kriging. A self-contained derivation of the proposed method is presented, and then it is verified by surrogate modeling test cases. The present method is integrated into a surrogatebased optimizer and tested for design optimizations. It is further demonstrated for inverse design of a transonic wing, parameterized with a number of design variables in the range from 36 to 108, using Reynolds-averaged Navier-Stokes flow and adjoint solvers. It is observed that, for the wing design with 36 and 54 variables, the weighted and conventional gradient-enhanced kriging are comparable, and both are much more efficient than kriging without using any gradient. For the wing design with 72 and 108 variables, the cost of training a gradient-enhanced kriging increases rapidly and becomes prohibitive. In contrast, the cost of training a weighted gradient-enhanced kriging is kept in an acceptable level, which makes it more practical for higher-dimensional problems.

show abstract

mentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

Weighted Gradient-Enhanced Kriging for High-Dimensional Surrogate Modeling and Design Optimization

et al. 2017

View full text Add to dashboard Cite

show abstract

“…A common approach to overcome the burden of high computational demand of complex evaluation functions during optimisation is the application of surrogate models (Forrester and Keane, 2009) to replace the original expensive objective functions by their computationally cheaper approximations to estimate fitness values of solutions. However, several issues may arise when applying surrogate-based optimisation to the airport ground movement problem.…”

Section: Introductionmentioning

confidence: 99%

A real-time Active Routing approach via a database for airport surface movement

Weiszer

Chen

Stewart

2015

Transportation Research Part C: Emerging Technologies

View full text Add to dashboard Cite

Airports face challenges due to the increasing volume of air traffic and tighter environmental restrictions which result in a need to actively integrate speed profiles into conventional routing and scheduling procedure. However, only until very recently, the research on airport ground movement has started to take into account such a speed profile optimisation problem actively so that not only time efficiency but also fuel saving and decrease in airport emissions can be achieved at the same time. It is envisioned that the realism of planning could also be improved through speed profiles. However, due to the multi-objective nature of the problem and complexity of the investigated models (objective functions), the existing speed profile optimisation approach features high computational demand and is not suitable for an on-line application. In order to make this approach more competitive for real-world application and to meet limits imposed by International Civil Aviation Organization for on-line decision time, this paper introduces a pre-computed database acting as a middleware to effectively separate the planning (routing and scheduling) module and the speed profile generation module. Employing a database not only circumvents duplicative optimisation for the same taxiway segments, but also completely avoids the computation of speed profiles during the on-line decision support owing a great deal to newly proposed database initialization procedures. Moreover, the added layer of database facilitates, in the future, more complex and realistic models to be considered in the speed profile generation module, without sacrificing on-line decision time. The experimental results carried out using data from a major European hub show that the proposed approach is promising in speeding up the search process.

show abstract

On the applicability of surrogate‐based Markov chain Monte Carlo‐Bayesian inversion to the Community Land Model: Case studies at flux tower sites

et al. 2016

View full text Add to dashboard Cite

Key Points:1. The feasibility of applying a Bayesian calibration technique to estimate CLM parameters is assessed; 2. CLM-simulated LH fluxes using the calibrated parameters are generally improved;3. The parameter values are likely transferable within the plant functional type. AbstractThe Community Land Model (CLM) has been widely used in climate and Earth system modeling. Accurate estimation of model parameters is needed for reliable model simulations and predictions under current and future conditions, respectively. In our previous work, a subset of hydrological parameters has been identified to have significant impact on surface energy fluxes at selected flux tower sites based on parameter screening and sensitivity analysis, which indicate that the parameters could potentially be estimated from surface flux observations at the towers. To date, such estimates do not exist.In this paper, we assess the feasibility of applying a Bayesian model calibration technique to estimate CLM parameters at selected flux tower sites under various site conditions. The parameters are estimated as a joint probability density function (PDF) that provides estimates of uncertainty of the parameters being inverted, conditional on climatologically-average latent heat fluxes derived from observations. We find that the simulated mean latent heat fluxes from CLM using the calibrated parameters are generally improved at all sites when compared to those obtained with CLM simulations using default parameter sets. Further, our calibration method also results in credibility bounds around the simulated mean fluxes which bracket the measured data. The modes (or maximum a posteriori values) and 95% credibility intervals of the site-specific posterior PDFs are tabulated as suggested parameter values for each site. Analysis of relationships between the posterior PDFs and site conditions suggests that the parameter values are likely correlated with the plant functional type, which needs to be confirmed in future studies by extending the approach to more sites.

show abstract

Recent advances in surrogate-based optimization

Cited by 1,815 publications

References 74 publications

Weighted Gradient-Enhanced Kriging for High-Dimensional Surrogate Modeling and Design Optimization

Weighted Gradient-Enhanced Kriging for High-Dimensional Surrogate Modeling and Design Optimization

A real-time Active Routing approach via a database for airport surface movement

On the applicability of surrogate‐based Markov chain Monte Carlo‐Bayesian inversion to the Community Land Model: Case studies at flux tower sites

Contact Info

Product

Resources

About