|
|
Evolution and Learning of Virtual Robots
Krčah, Peter ; Mráz, František (advisor) ; Kvasnička, Vladimír (referee) ; Pilát, Martin (referee)
Title: Evolution and Learning of Virtual Robots Author: RNDr. Peter Krčah Department: Department of Software and Computer Science Education Supervisor: RNDr. František Mráz, CSc. Abstract: Evolutionary robotics uses evolutionary algorithms to automatically design both body and controller of a robot. We describe two contributions to automated design of virtual robotic creatures. First, we introduce a nature-inspired method that allows virtual robots to modify their morphology through lifetime learning. We show that such morphological plasticity makes it possible to evolve robots that can dynamically adjust their morphology to the environment they are placed into. We also show that by reshaping the fitness landscape, learning reduces computation cost required to evolve a robot with a given target fitness even in a single environment. In the second contribution, we show that for certain problems in evolutionary robotics, premature convergence to local optima can be avoided by ignoring the original objective and searching for any novel behaviors instead (a technique known as Novelty Search). Keywords: Evolution of Virtual Creatures, Body-brain Coevolution, Morphological Plasticity, Neural Networks, Learning
|
|
|
Phytoplankton of the Tatra lakes: factors influencing taxonomic composition and biomass
Červenková, Lenka ; Nedbalová, Linda (advisor) ; Štefková, Elena (referee)
The aim of this diploma thesis was to describe species composition and biomass of phytoplankton in the Tatra Mountains lakes (Slovakia, Poland) and environmental factors influencing them. The surface layer of 89 lakes was sampled in September 2004. Flagellates from Chrysophyta, Dinophyta and Cryptophyta most frequently dominated the phytoplankton biomass. Lake in the same valley usually differed in their dominant taxonomic group as a result of different lake morphometry, catchment type and detailed geological structure. Based on the level of phytoplankton biovolume, the majority of lakes were ultraoligotrophic and oligotrophic, however, mesotrophic and even eutrophic conditions were recorded in some forest and meadow lakes. An allometric relationship between phytoplankton biomass and chlorophyll a was revealed. Specific chlorophyll content did not differ among taxonomic groups. Altogether 233 species were determined, the most diverse group were Chlorophyta. Species richness of lake was 3-46 and it was negatively correlated with altitude and positively correlated with lake area. The majority of species were rare both in term of biomass and number of lakes occupied by a particular species. Average local species biomass was positively correlated with regional distribution. The similarity of species...
|
|
|
Plant perception of soil heterogeneity in the field
Hrouda, Adam ; Weiser, Martin (advisor) ; Martínková, Jana (referee)
Nutrients are usually patchily distributed in natural soils. Plants are often able to respond to nutrient heterogeneity in artificial conditions by active plastic changes of root system morphology. The occurrence or magnitude of a foraging response can be altered by the presence of competition. However, it is unclear to what extent root foraging takes place in the field. I conducted a field experiment in order to determine the effect of an artificial nutrient patch on fine belowground biomass of (a) an established community and (b) model plants. The study array consisted of a grid of 30×30 cm plots with model plants located in the centre. Half of the plots contained the artificial patch located 5.5 cm from the model plant. Fertilizer patch treatment did not increase mean plot fine underground biomass. Instead, fine underground biomass was higher in places of greater soil moisture estimated from mean plot EIVs. Neither total model plant root biomass nor proportion of roots in the enriched quarter increased in the fertilizer treatment. Competition was probably higher in fertilized than in control plots judging by a 2-fold increase in death rate of model plants. However, greater proportion of model plants flowered in the treatment plots. Possible causes include a plastic response to the patch as well...
|
|
|
Evolution and Learning of Virtual Robots
Krčah, Peter ; Mráz, František (advisor) ; Kvasnička, Vladimír (referee) ; Pilát, Martin (referee)
Title: Evolution and Learning of Virtual Robots Author: RNDr. Peter Krčah Department: Department of Software and Computer Science Education Supervisor: RNDr. František Mráz, CSc. Abstract: Evolutionary robotics uses evolutionary algorithms to automatically design both body and controller of a robot. We describe two contributions to automated design of virtual robotic creatures. First, we introduce a nature-inspired method that allows virtual robots to modify their morphology through lifetime learning. We show that such morphological plasticity makes it possible to evolve robots that can dynamically adjust their morphology to the environment they are placed into. We also show that by reshaping the fitness landscape, learning reduces computation cost required to evolve a robot with a given target fitness even in a single environment. In the second contribution, we show that for certain problems in evolutionary robotics, premature convergence to local optima can be avoided by ignoring the original objective and searching for any novel behaviors instead (a technique known as Novelty Search). Keywords: Evolution of Virtual Creatures, Body-brain Coevolution, Morphological Plasticity, Neural Networks, Learning
|
|
|
Clonal integration of Agrostis stolonifera in heterogeneous soil environment
Duchoslavová, Jana ; Weiser, Martin (advisor) ; Klimešová, Jitka (referee)
Clonal plants may be able to cope with spatial heterogeneity due to the physiological integration of ramets. Previous studies demonstrated that benefits of clonal integration increase with patch contrast between individual ramets. However, the same magnitude of contrast may be perceived differently in rich and poor environments. According to the theoretical work of Caraco and Kelly (1991), I expected these benefits to be the greatest in overall poor conditions and high between-patch contrast. To test this hypothesis, I conducted experiments with pairs of ramets of a stoloniferous grass, Agrostis stolonifera, grown in variously nutrient rich conditions. The experiment with pairs of ramet of similar developmental age showed only very weak effect of integration on growth of ramets, although integration significantly improved survival of ramets and also affected root-shoot ratio of ramets. Nevertheless, there were considerable benefits of integration in the experiment with developmentally older mother ramets and their daughter ramets. Contrary to the predictions, the benefits of integration were bigger in rich conditions and they decreased with increasing between-patch contrast. In addition, effect of integration on root-shoot ratio of ramets was opposite to the expected specialization for acquisition...
|
|
|
Phytoplankton of the Tatra lakes: factors influencing taxonomic composition and biomass
Červenková, Lenka ; Nedbalová, Linda (advisor) ; Štefková, Elena (referee)
The aim of this diploma thesis was to describe species composition and biomass of phytoplankton in the Tatra Mountains lakes (Slovakia, Poland) and environmental factors influencing them. The surface layer of 89 lakes was sampled in September 2004. Flagellates from Chrysophyta, Dinophyta and Cryptophyta most frequently dominated the phytoplankton biomass. Lake in the same valley usually differed in their dominant taxonomic group as a result of different lake morphometry, catchment type and detailed geological structure. Based on the level of phytoplankton biovolume, the majority of lakes were ultraoligotrophic and oligotrophic, however, mesotrophic and even eutrophic conditions were recorded in some forest and meadow lakes. An allometric relationship between phytoplankton biomass and chlorophyll a was revealed. Specific chlorophyll content did not differ among taxonomic groups. Altogether 233 species were determined, the most diverse group were Chlorophyta. Species richness of lake was 3-46 and it was negatively correlated with altitude and positively correlated with lake area. The majority of species were rare both in term of biomass and number of lakes occupied by a particular species. Average local species biomass was positively correlated with regional distribution. The similarity of species...
|
guest ::
