Single- and multiobjective evolutionary optimization assisted by Gaussian random field metamodels

Emmerich, Michael; Giannakoglou, Kyriakos C.; Naujoks, Boris

doi:10.1109/tevc.2005.859463

Cited by 624 publications

(438 citation statements)

References 26 publications

Supporting

Mentioning

427

Contrasting

Unclassified

Order By: Relevance

“…Although recognized as less efficient compared to the EI in the case of deterministic experiments [19], this method seems worth studying in this benchmark because it does not need any modification to handle noise. It also has shown successful applications in kriging-based multiobjective optimization [8,27].…”

Section: Minimal Quantile Criteria (Mq)mentioning

confidence: 99%

A benchmark of kriging-based infill criteria for noisy optimization

Picheny

Wagner

Ginsbourger³

2013

Struct Multidisc Optim

270

178

View full text Add to dashboard Cite

Responses of many real-world problems can only be evaluated perturbed by noise. In order to make an efficient optimization of these problems possible, intelligent optimization strategies successfully coping with noisy evaluations are required. In this article, a comprehensive review of existing kriging-based methods for the optimization of noisy functions is provided. In summary, ten methods for choosing the sequential samples are described using a unified formalism. They are compared on analytical benchmark problems, whereby the usual assumption of homoscedastic Gaussian noise made in the underlying models is meet. Different problem configurations (noise level, maximum number of observations, initial number of observations) and setups (covariance functions, budget, initial sample size) are considered. It is found that the choices of the initial sample size and the covariance function are not critical. The choice of the method, however, can result in significant differences in the performance. In particular, the three most intuitive criteria are found as poor alternatives. Although no criterion is found consistently more efficient than the others, two specialized methods appear more robust on average.

show abstract

Section: Minimal Quantile Criteria (Mq)mentioning

confidence: 99%

A benchmark of kriging-based infill criteria for noisy optimization

Picheny

Wagner

Ginsbourger³

2013

Struct Multidisc Optim

270

178

View full text Add to dashboard Cite

show abstract

“…We take a traditional way to do the job, i.e., using model as a prescreening mechanism [5]. Similar to ParEGO, at every iteration, MOEA/D runs on a learned model, then the best result is selected for evaluation.…”

Section: Moea/d With Gp Model Assistantmentioning

confidence: 99%

“…However, because of the di erence between single and multiple objectives problem solving, those methods cannot be readily applied in MOEA. Recent researches on MOEA with metamodel assistance also appeared, notably in [11] and [5], which we think represent two kinds of methods to transform metamodel assistance from single objective to multiple objectives.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

On the Performance of Metamodel Assisted MOEA/D

Liu

Zhang

Tsang

et al.

Advances in Computation and Intelligence

View full text Add to dashboard Cite

Abstract. MOEA/D is a novel and successful Multi-Objective Evolutionary Algorithms(MOEA) which utilizes the idea of problem decomposition to tackle the complexity from multiple objectives. It shows better performance than most nowadays mainstream MOEA methods in various test problems, especially on the quality of solution's distribution in the Pareto set. This paper aims to bring the strength of metamodel into MOEA/D to help the solving of expensive black-box multiobjective problems. Gaussian Random Field Metamodel(GRFM) is chosen as the approximation method. The performance is analyzed and compared on several test problems, which shows a promising perspective on this method.

show abstract

“…Statistical approaches such as Design of Experiments (DoE) from Taguchi type models to sophisticated Kriging are typical. Emmerich (2005) describes the development of robust algorithms for optimization with time-consuming evaluations. The main working principle of these techniques is to combine spatial interpolation techniques with EAs.…”

Section: Expensive Function Evaluationsmentioning

confidence: 99%

An Overview on Evolutionary Algorithms

Data Mining Using Grammar Based Genetic Programming and Applications

View full text Add to dashboard Cite

Single- and multiobjective evolutionary optimization assisted by Gaussian random field metamodels

Cited by 624 publications

References 26 publications

A benchmark of kriging-based infill criteria for noisy optimization

A benchmark of kriging-based infill criteria for noisy optimization

On the Performance of Metamodel Assisted MOEA/D

An Overview on Evolutionary Algorithms

Contact Info

Product

Resources

About