%0 Conference Paper %B Genetic and Evolutionary Computation Conference Companion (GECCO '22) %D 2022 %T Diversification Techniques and Distance Measures in Evolutionary Design of 3D Structures %A Adam Klejda %A Maciej Komosinski %A Agnieszka Mensfelt %X Evolutionary algorithms are among the most successful metaheuristics for hard optimization problems. Nonetheless, there is still much room for improvement of their effectiveness, especially in the multimodal problems, where the algorithms are prone to falling into unsatisfactory local optima. One of the solutions to this problem may be to encourage a broader exploration of the solution space. Motivated by this premise, we compare the evolutionary algorithm without niching, with niching, the novelty search, and the two-criteria optimization (NSGA-II) where the criteria of fitness and diversity are not aggregated. We investigate these methods in the context of automated design of three-dimensional structures, which is one of the hardest optimization problems, often characterized by a rugged fitness landscape arising from a complex genotype to phenotype mapping. In the experiments we optimize 3D structures towards two different goals, height and velocity, using two genetic encodings and three distance measures: two phenetic ones and a genetic one. We demonstrate how different distance measures and diversity promotion mechanisms influence the fitness of the obtained solutions. %B Genetic and Evolutionary Computation Conference Companion (GECCO '22) %I ACM %C Boston, USA %G eng %R 10.1145/3520304.3528948 %0 Report %D 2020 %T Human perception of similarity of 3D graph structures %A Maciej Komosinski %A Agnieszka Mensfelt %X This report describes the study of how humans perceive similarity of simple three-dimensional graph structures. Participants of this study were required to align pairs of 3D structures the best they could, then match all vertices of these structures, evaluate their perceived similarity on a numerical scale, and justify their decisions as a textual response. The outcomes of this process were analyzed and compared to the outcomes of a heuristic computer algorithm that maximized the alignment of pairs of 3D structures and matched their vertices. The influence of personal characteristics of participants such as their gender, age, handedness, education, but also time required to complete each task, on the quality of the matching of vertices was evaluated. The consistency of human responses was also verified. The participants turned out to be more consistent (both between themselves and with the algorithm) in the degree of similarity estimated than in matching of vertices. Personal characteristics of the subjects did not have an influence on their similarity assessments. %G eng %0 Journal Article %J Journal of Computational Science %D 2017 %T Multi-agent simulation of benthic foraminifera response to annual variability of feeding fluxes %A Maciej Komosinski %A Agnieszka Mensfelt %A Jarosław Tyszka %A Jan Goleń %X In this work we describe a novel simulation model of foraminifera and their microhabitat. The simulations reported here are focused on the response of foraminiferal populations to environmental feeding fluxes. The experiments allowed to calibrate the model and to simulate realistic population patterns known from culture experiments, as well as from oceanographic and paleoecologic studies. Variability of annual food flux has a direct impact on productivity of foraminifera: population sizes closely follow the intensity of constant and seasonal food fluxes in both scenarios. This correlation between the food influx and population size is interpreted as the consequence of changing the carrying capacity of the system. Seasonal pulses of particulate organic matter enhance the population size which is represented by a higher number of fossilized shells. Our model offers a flexible experimental design to run sophisticated in silico experiments. This approach reveals a novel methodology for testing sensitivity of fossil and recent foraminiferal assemblages to environmental changes. Furthermore, it facilitates predictive applications for monitoring studies based on simulation of various scenarios. %B Journal of Computational Science %V 21 %P 419–431 %U http://www.framsticks.com/files/common/SimulationForaminiferaFeedingFluxes.pdf %R 10.1016/j.jocs.2016.09.009 %0 Book Section %B Man–Machine Interactions 4 %D 2016 %T Application of a morphological similarity measure to the analysis of shell morphogenesis in Foraminifera %A Maciej Komosinski %A Agnieszka Mensfelt %A Topa, Paweł %A Jarosław Tyszka %E Gruca, Aleksandra %E Brachman, Agnieszka %E Kozielski, Stanisław %E Czachórski, Tadeusz %X This work evaluates the genotype-to-phenotype mapping defined by one of the models of growth of foraminifera. Foraminifera are simple unicellular organisms with very diverse morphologies. To analyze the mapping, a morphological similarity measure is needed that compares 3D structures. One of the key components of the similarity estimation algorithm is Singular Value Decomposition (SVD). Since this algorithm is heavily used and its performance is important, four SVD implementations have been compared in this work. Distance matrices of the phenotypes obtained for equally distant genotypes were computed using the similarity measure. For the visualization of the phenotype space, multidimensional scaling techniques were used. Visual comparison of the genotype and the phenotype spaces revealed characteristics and potential weaknesses of the analyzed model of foraminifera growth, and demonstrated usefulness of the proposed approach. %B Man–Machine Interactions 4 %S Advances in Intelligent Systems and Computing %I Springer %V 391 %P 215–224 %@ 978-3-319-23436-6 %U http://www.framsticks.com/files/common/ForaminiferaGenotypePhenotypeMapping.pdf %R 10.1007/978-3-319-23437-3_18 %0 Generic %D 2014 %T Foraminifera: genetics, morphology, simulation, evolution %A Maciej Komosinski %A Agnieszka Mensfelt %A Topa, Paweł %A Jarosław Tyszka %A Szymon Ulatowski %U http://www.framsticks.com/foraminifera