Multilevel optimization strategies based on metamodel-assisted evolutionary algorithms, for computationally expensive problems

Evolutionary Optimization with Dynamic Fidelity Computational Models

“…Subsequently, the obtained solution is then refined by space mapping. On the other hand, instances of deterministic adaptation can be found in [4,8,10]. In [4], the evolutionary search employs computational models of different grid mesh that varies from coarse to fine grids at pre-defined stages of the GA search.…”

“…In [4], the evolutionary search employs computational models of different grid mesh that varies from coarse to fine grids at pre-defined stages of the GA search. In a multi-island GA optimizer [8,10], multiple subpopulation of individuals are equipped with models of variable fidelity levels. Individuals then migrate across subpopulation which are subsequently evaluated based on the computational model of the respective subpopulation.…”

“…A common assumption is that higher fidelity models are generally more accurate at the expense of a higher computational cost. The complexity and level of details of a physical system comes in many forms, from variable mathematical models [7], variable parametric formulations [4,[8][9][10], variable operating conditions [11] to variable residual tolerance levels [12].…”

Distributed evolutionary algorithms with hierarchical evaluation

“…Taking this into consideration, herein a generalized (µ, λ)EA (with µ parents and λ offspring) implementing standard evolution operators, will be used as the core search engine. Structuring the evolutionary search in hierarchical manner is a means to reduce the overall CPU cost for carrying out the optimization, combining different tools (Kampolis et al 2007, Kampolis andGiannakoglou 2008). The gain from using hierarchical search is superimposed to that expected from the use of a "better" EA, "better" evolution operators and so forth.…”

“…The gain from using hierarchical search is superimposed to that expected from the use of a "better" EA, "better" evolution operators and so forth. According to the literature survey, existing hierarchical schemes can be classified to algorithms relying on different evaluation methods to computer the fitness or cost function of candidate solutions (with different fidelity and CPU cost) (Eby et al 1998, Herrera et al 1999, Sefrioui and Périaux 2000, Karakasis et al 2007, Kampolis et al 2007, Kampolis and Giannakoglou 2008, different search techniques (Muyl et al 2004, Poloni et al 2000, Knowles and Corne 2000 and different chromosome sizes (Lin et al 1994) or numbers of design variables (Désidéri andJanka 2003, Duvigneau et al 2006). The present paper is concerned with the first class of hierarchical methods, i.e.…”