Multiobjective Optimization on a Limited Budget of Evaluations Using Model-Assisted -Metric Selection

By Wolfgang Ponweiser, Tobias Wagner, Dirk Biermann, Markus Vincze. Parallel Problem Solving from Nature, 2008

303 citations recorded by the source corpus.

Subfields: Computational Theory and Mathematics, Management Science and Operations Research, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/9fdc5b56f5af2bbe07f38e395a2ea8c01783bc06

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Theory and Mathematics

In Management Science and Operations Research

In Statistics and Probability

Citations in the corpus

Kriging is well-suited to parallelize optimization (2010, unknown): “A representative example is given by the EGO algorithm [22], a sequential black-box optimization procedure, which has gained popularity during the last decade and inspired numerous recent works in the ﬁeld [10, 17, 18, 19, 20, 26, 28, 36, 44, 50].”
Model-Based Multi-objective Optimization: Taxonomy, Multi-Point Proposal, Toolbox and Benchmark (2015, methodology): “SMSEGO [18] LHS Models for each objective Single-objective optimization of the hypervolume contribution One point, same criterion Total budget”
Differentiable Expected Hypervolume Improvement for Parallel Multi-Objective Bayesian Optimization (2020, methodology): “2 qPAREGO Previous work has only considered unconstrained sequential optimization with ParEGO [40, 7] and ParEGO is often optimized with gradient-free methods [53].”
A Classification-Based Surrogate-Assisted Evolutionary Algorithm for Expensive Many-Objective Optimization (2019, methodology): “In SMS-EGO, the fitness of an individual is defined as the contribution of this individual to the HV score of the population, and a Kriging model is built to approximate the function of the HV [51].”
PlatEMO: A MATLAB Platform for Evolutionary Multi-Objective Optimization [Educational Forum] (2017, unknown): “SmS-Ego [73] 2008 S-mETRiC-SElECTion-bASEd EffiCiEnT globAl oPTimiZATion”
A benchmark of kriging-based infill criteria for noisy optimization (2013, methodology): “It also has shown successful applications in kriging-based multiobjective optimization [8, 27].”
Surrogate-Assisted Cooperative Swarm Optimization of High-Dimensional Expensive Problems (2017, methodology): “The most commonly used surrogate models include polynomial regression (PR) [19], also known as response surface methodology [19], support vector machines (SVMs) [20], [21], [22], artificial neural networks (ANNs) [12], [23], [24], radial basis function (RBF) networks [25], [26], [27], [28], [29], an”
Recent Advances in Bayesian Optimization (2022, methodology): “In SMS-EGO, a Kriging model is built for each objective, then HV is optimized to select new samples, where the LCB is adopted to calculate the fitness values.”
A Surrogate-Assisted Reference Vector Guided Evolutionary Algorithm for Computationally Expensive Many-Objective Optimization (2018, methodology): “only three algorithms [6], [38], [41] have been tested on multiobjective benchmark problems with more than three objectives.”
Bayesian Optimization for Adaptive Experimental Design: A Review (2020, methodology): “The method performs better than [107], but suffers in high dimensions and can be computationally expensive.”
Review of surrogate modeling in water resources (2012, methodology): “…by a weighting scheme) to be approximated by a single response surface surrogate [e.g., di Pierro et al., 2009; Knowles, 2006; Zhang et al., 2010], and (3) using multiple surrogate models for multiple objectives [e.g., Baú and Mayer, 2006; Keane, 2006; Li et al., 2008; Ponweiser et al., 2008].”
Expensive Multiobjective Optimization by MOEA/D With Gaussian Process Model (2010, methodology): “Clearly, MOEA/D-EGO with single model is much Instance Number of Evaluations ParEGO MOEA/D-EGO Lowest Mean Std Lowest Mean Std ZDT1 200 0.0513 0.0945 0.0316 0.0120 0.0148 0.0016 ZDT2 200 0.0440 0.0864 0.0483 0.0121 0.0156 0.0022 ZDT3 200 0.2071 0.3818 0.1385 0.0435 0.0665 0.0189 ZDT4 200 2.9322 6.23”
A Survey of Multiobjective Evolutionary Algorithms Based on Decomposition (2017, result): “) demonstrated that when limited number of function evaluations are allowed, the performance of MOEA/D-EGO is comparable to that of ParEGO [76] and SMS-EGO [77].”