Evolutionary approaches to solve three challenging engineering tasks Original Research Article

hree applications of evolutionary algorithms, namely the optimization of the nuclear core reload design, the synthesis of multi-layer optical coatings, and the synthesis of chemical engineering plants, are presented in this paper. The examples demonstrate the applicability of the evolutionary approach to solve complex real-world problems. These problems are often mixed-integer, variable-dimensional and multi-criteria optimization problems. Additionally, instead of an objective function, given in a closed form, complex simulation models are used as an objective function, thus prohibiting success by means of classical analysis. Although standard evolutionary algorithms are not able to solve the three complex tasks presented here, enhanced evolutionary algorithms (EAs) clearly demonstrate their potential to do so. Through the three examples, we show the necessary ingredients for tackling real-world optimization problems. In addition to adequate representations and appropriate evolutionary operators, one needs to integrate expert knowledge and heuristic operators to guide the search. Finally, parallel EAs are needed to improve the reliability and speed up of evolutionary algorithms.