ISBN 978-3-8439-5120-3

978-3-8439-5120-3, Reihe Ingenieurwissenschaften

Christian Kasten
Optimizing evolutionary algorithms with gender and application to the modelling of turbulent combustion

143 Seiten, Dissertation Universität der Bundeswehr München (2022), Softcover, B5

Zusammenfassung / Abstract

Evolutionary algorithm have been extensively used for numerous optimization problems with great success in the past years. Due to the high cost of multidimensional complex optimization problems, it is of high interest to increase the efficiency of Evolutionary algorithms and to reach better solutions with less computational effort. Most of the aspects that have been proven to be effective in nature are already transferred to Evolutionary algorithm. But one specific concept is not present in today’s Evolutionary algorithm codes and that is the concept of gender using a characteristic of an individual.

After a detailed introduction of gene expression programming a new strategy to introduce gender using an individual’s characteristic is suggested in this work. The new strategy is tested on three different benchmark problems with increasing complexity. In addition, the effectiveness and efficiency of the new method is compared to the original one. It is shown that the new method outperforms the original one in terms of success rate and success speed.

To ensure this method is ready to be used in real engineering problems, it is also applied and compared on two more complicated problems from the field of Computational Fluid Dynamics. A Direct Numerical Simulation (DNS) database is introduced and GEP with gender is applied to closing the filtered reaction rate term as well as the subgrid scale scalar dissipation rate in the context of Large Eddy Simulation (LES) of turbulent premixed combustion.