ParEGO: a hybrid algorithm with on-line landscape approximation for expensive multiobjective optimization problems

By Joshua D. Knowles. IEEE Transactions on Evolutionary Computation, 2006

1,291 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computational Theory and Mathematics

View Paper: https://www.semanticscholar.org/paper/73b5b196b35fb23e1f908d73b787c2c2942fadb5

How researchers see this paper

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

In Artificial Intelligence

In Computational Theory and Mathematics

Citations in the corpus

Offline Data-Driven Evolutionary Optimization Using Selective Surrogate Ensembles (2019, background): “This assumption does not hold, unfortunately, for many real-world optimization problems, where either computationally intensive numerical simulations or expensive experiments must be performed for ﬁtness evaluations [3], [4].”
A Classification-Based Surrogate-Assisted Evolutionary Algorithm for Expensive Many-Objective Optimization (2019, methodology): “4) Specific Parameter Setting in Each Algorithm: For ParEGO, the number of weight vectors is set to 11 for biobjective problems and 15 for tri-objective problems, and the maximum number of surrogate-assisted fitness approximation before the surrogate update is set to 200 000 as recommended in [37].”
A Tutorial on Bayesian Optimization (2018, background): “…in that same literature include multi-ﬁdelity optimization (Huang et al., 2006; S´obester et al., 2004), multi-objective optimization (Keane, 2006; Knowles, 2006; Moˇckus and Moˇckus, 1991), and a study of convergence rates (Calvin, 1997; Calvin and ˇZilinskas, 2000; Calvin and ˇZilinskas,…”
A Surrogate-Assisted Reference Vector Guided Evolutionary Algorithm for Computationally Expensive Many-Objective Optimization (2018, methodology): “1) Number of individuals to be evaluated using the original objective functions in the initialization phase (from [28] and [51]) = maximum number of individuals in the training archive, NI = 11n − 1.”
Surrogate-assisted evolutionary computation: Recent advances and future challenges (2011, methodology): “Due to this property, Gaussian processes have increasingly been employed as surrogates in evolutionary single-and multi-objective optimization [41] [42] [43] [44].”
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”
Response surface approximation of pareto optimal front in multi-objective optimization (2004, background): “Some of the latest ones include strength Pareto evolutionary algorithm (Zitzler and Thiele) [6], Pareto archived evolutionary strategies (PAES by Knowles and Corne) [7], elitist nondominated sorting genetic algorithm (NSGA-II by Deb et al.)”
Kriging is well-suited to parallelize optimization (2010, neutral): “Moreover, [26] proposes an adaptation to multi-objective optimization, [17] proposes an original multi-objective adaptation of EGO for physical experiments, and [28] focuses on robust criteria for multiobjective constrained optimization with applications to laminating processes.”
A Random Forest-Assisted Evolutionary Algorithm for Data-Driven Constrained Multiobjective Combinatorial Optimization of Trauma Systems (2020, background): “SAEAs for single-objective optimization [6], [25], [29], [30]; multiobjective optimization [12], [22], [31], [32]; and manyobjective optimization [13], [33] have been proposed.”
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].”
Multiobjective optimization techniques applied to engineering problems (2010, methodology): “Knowles (2005) uses multiobjective optimization in scenarios where the evaluation of each solution is financially and/or temporally expensive.”
Model-Based Multi-objective Optimization: Taxonomy, Multi-Point Proposal, Toolbox and Benchmark (2015, methodology): “Early work in the multi-objective scenario has either scalarized the objectives [15] to allow EGO to be directly used or has optimized EGO’s figure of merit for different models in parallel using MOEA [11,19].”
Bayesian Optimization is Superior to Random Search for Machine Learning Hyperparameter Tuning: Analysis of the Black-Box Optimization Challenge 2020 (2021, background): “Additionally, while we only focused on unconstrained single-objective optimization performance, many relevant problems have additional black-box constraints that need to be satisﬁed or are more naturally phrased as multi-objective optimization problems [12, 13, 21, 32, 40].”
Recent advances and applications of surrogate models for finite element method computations: a review (2022, methodology): “Of the algorithm that are more widely used are the multiobjective genetic algorithm NSGA-II (Deb et al. 2002), the ParEGO algorithm (Knowles 2006), and the RASM method (Loshchilov et al. 2010).”
Bayesian Optimization for Adaptive Experimental Design: A Review (2020, methodology): “Many methods have been proposed for using Bayesian optimisation for multi-objective optimisation [106]–[109], but these suffer from computational limitations because the acquisition function generally requires computation for all objective functions and as the number of objective functions”
Differentiable Expected Hypervolume Improvement for Parallel Multi-Objective Bayesian Optimization (2020, methodology): “For example, ParEGO [39] randomly scalarizes the objectives and uses Expected Improvement [37], and SMS-EGO [50] uses HV in a UCB-based acquisition function and is more scalable than EHVI [51].”