Generalizing Surrogate-Assisted Evolutionary Computation

By Dudy Lim, Yaochu Jin, Yew-Soon Ong, Bernhard Sendhoff. IEEE Transactions on Evolutionary Computation, 2010

445 citations recorded by the source corpus.

Using surrogate models in evolutionary search provides an efficient means of handling today's complex applications plagued with increasing high-computational needs. Recent surrogate-assisted evolutionary frameworks have relied on the use of a variety of different modeling approaches to approximate the complex problem landscape. From these recent studies, one main research issue is with the choice of modeling scheme used, which has been found to affect the performance of evolutionary search significantly. Given that theoretical knowledge available for making a decision on an approximation model a priori is very much limited, this paper describes a generalization of surrogate-assisted evolutionary frameworks for optimization of problems with objectives and constraints that are computationally expensive to evaluate. The generalized evolutionary framework unifies diverse surrogate models synergistically in the evolutionary search. In particular, it focuses on attaining reliable search performance in the surrogate-assisted evolutionary framework by working on two major issues: 1) to mitigate the 'curse of uncertainty' robustly, and 2) to benefit from the 'bless of uncertainty.' The backbone of the generalized framework is a surrogate-assisted memetic algorithm that conducts simultaneous local searches using ensemble and smoothing surrogate models, with the aims of generating reliable fitness prediction and search improvements simultaneously. Empirical study on commonly used optimization benchmark problems indicates that the generalized framework is capable of attaining reliable, high quality, and efficient performance under a limited computational budget.

Subfields: Artificial Intelligence, Computational Theory and Mathematics

View Paper: https://www.semanticscholar.org/paper/3e6ff64046681968f5bf23582b8e36e5558112d8

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

Data-Driven Evolutionary Optimization: An Overview and Case Studies (2019, background): “as shown in [47, 61], an EA may benefit from the approximation errors introduced by surrogates, and therefore, it is essential in SAEAs to make full use of the limited data.”
Surrogate-assisted evolutionary computation: Recent advances and future challenges (2011, background): “The idea of taking advantage of approximation errors introducedby surrogateswas further exploited in [64].”
A social learning particle swarm optimization algorithm for scalable optimization (2015, background): “In real-world applications, the time cost of fitness evaluations is problem-dependent [38,50], which is beyond the scope of this work.”
Offline Data-Driven Evolutionary Optimization Using Selective Surrogate Ensembles (2019, methodology): “Another idea is to use a combination of global and local surrogate models in which the global model is used to smoothen out the local optimums while the local ones are utilized for exploiting the local details of the ﬁtness landscape [27]–[29].”
On Using Surrogates with Genetic Programming (2015, methodology): “…the surrogate model and generations that are evaluated using the expensive ﬁtness function (Rattle, 1998; Jin et al., 2002), to use surrogate models for supporting local search (Lim et al., 2010), or for pre-selection amongst individuals generated by a genetic operator (Em-merich et al., 2002).”
A Gaussian Process Surrogate Model Assisted Evolutionary Algorithm for Medium Scale Expensive Optimization Problems (2014, methodology): “Experimental studies on a set of widely used test instances have shown that comparable results can be obtained using 12% to 50% expensive function evaluations compared with three other state-of-the-art SAEAs.”
Memetic algorithms and memetic computing optimization: A literature review (2012, methodology): “[110] have recently proposed a framework composed of an ensemble of approximation models as well smoothing models.”
Resampling Methods for Meta-Model Validation with Recommendations for Evolutionary Computation (2012, background): “In addition, hierarchical models (Kampolis et al., 2007; Zhou et al., 2007; Lim et al., 2010) which pick up the meta-modeling approach and expand it to multiple models of different complexity are a straightforward approach and a matter of current research.”