%0 Conference Paper %B Genetic and Evolutionary Computation Conference Companion (GECCO '22) %D 2022 %T Fitness Diversification in the Service of Fitness Optimization: a Comparison Study %A Kamil Basiukajc %A Maciej Komosinski %A Konrad Miazga %X Blindly chasing after fitness is not the best strategy for optimization of hard problems, as it usually leads to premature convergence and getting stuck in low-quality local optima. Several techniques such as niching or quality-diversity algorithms have been established that aim to alleviate the selective pressure present in evolutionary algorithms and to allow for greater exploration. Yet another group of methods which can be used for that purpose are fitness diversity methods. In this work we compare the standard single-population evolution against three fitness diversity methods: fitness uniform selection scheme (FUSS), fitness uniform deletion scheme (FUDS), and convection selection (ConvSel). We compare these methods on both mathematical and evolutionary design benchmarks over multiple parametrizations. We find that given the same computation time, fitness diversity methods regularly surpass the performance of the standard single-population evolutionary algorithm. %B Genetic and Evolutionary Computation Conference Companion (GECCO '22) %I ACM %C Boston, USA %G eng %U http://www.framsticks.com/files/common/FitnessDiversity.pdf %R 10.1145/3520304.3528949 %0 Journal Article %J Foundations of Computing and Decision Sciences %D 2019 %T Mappism: formalizing classical and artificial life views on mind and consciousness %A Iwo Błądek %A Maciej Komosinski %A Konrad Miazga %B Foundations of Computing and Decision Sciences %V 44 %P 55–99 %G eng %U http://www.framsticks.com/files/common/MappismConsciousness.pdf %R 10.2478/fcds-2019-0005 %0 Book Section %B Artificial Life Conference Proceedings %D 2019 %T Measuring properties of movement in populations of evolved 3D agents %A Maciej Komosinski %A Konrad Miazga %E Harold Fellermann %E Jaume Bacardit %E Angel Goni-Moreno %E Rudolf M. Füchslin %B Artificial Life Conference Proceedings %I MIT Press %P 485–492 %G eng %R 10.1162/isal_a_00208 %0 Book Section %B Genetic and Evolutionary Computation Conference (GECCO '19), July 13–17, 2019, Prague, Czech Republic %D 2019 %T Parametrizing Convection Selection: Conclusions from the Analysis of Performance in the NKq Model %A Maciej Komosinski %A Konrad Miazga %B Genetic and Evolutionary Computation Conference (GECCO '19), July 13–17, 2019, Prague, Czech Republic %I ACM %P 804–811 %G eng %U http://www.framsticks.com/files/common/ConvectionSelectionNKqModel.pdf %R 10.1145/3321707.3321864 %0 Report %D 2019 %T Properties of movement of 3D agents %A Krzysztof Gorgolewski %A Maciej Komosinski %A Konrad Miazga %A Krzysztof Rosinski %A Paweł Rychły %I Poznan University of Technology, Institute of Computing Science %G eng %U http://www.framsticks.com/files/common/PropertiesOfMovementOf3DAgents.pdf %0 Journal Article %J Journal of Computational Science %D 2018 %T Comparison of the tournament-based convection selection with the island model in evolutionary algorithms %A Maciej Komosinski %A Konrad Miazga %X Convection selection is an approach to multipopulational evolutionary algorithms where solutions are assigned to subpopulations based on their fitness values. Although it is known that convection selection can allow the algorithm to find better solutions than it would be possible with a standard single population, the convection approach was not yet compared to other, commonly used architectures of multipopulational evolutionary algorithms, such as the island model. In this paper we describe results of experiments which facilitate such a comparison, including extensive multi-parameter analyses. We show that approaches based on convection selection can obtain better results than the island model, especially for difficult optimization problems such as those existing in the area of evolutionary design. We also introduce and test a generalization of the convection selection which allows for adjustable overlapping of fitness ranges of subpopulations; the amount of overlapping influences the exploration vs. exploitation balance. %B Journal of Computational Science %V 32 %P 106–114 %G eng %U http://www.framsticks.com/files/common/ConvectionSelectionVsIslandModel.pdf %R 10.1016/j.jocs.2018.10.001 %0 Journal Article %J PPAM 2017 proceedings, Lecture Notes in Computer Science %D 2018 %T Tournament-based convection selection in evolutionary algorithms %A Maciej Komosinski %A Konrad Miazga %X One of the problems that single-threaded (non-parallel) evolutionary algorithms encounter is premature convergence and the lack of diversity in the population. To counteract this problem and improve the performance of evolutionary algorithms in terms of the quality of optimized solutions, a new subpopulation-based selection scheme - the convection selection - is introduced and analyzed in this work. This new selection scheme is compared against traditional selection of individuals in a single-population evolutionary processes. The experimental results indicate that the use of subpopulations with fitness-based assignment of individuals yields better results than both random assignment and a traditional, non-parallel evolutionary architecture. %B PPAM 2017 proceedings, Lecture Notes in Computer Science %V 10778 %P 466–475 %G eng %U http://www.framsticks.com/files/common/TournamentBasedConvectionSelectionEvolutionary.pdf %R 10.1007/978-3-319-78054-2_44