Refine
H-BRS Bibliography
- yes (14)
Departments, institutes and facilities
- Institut für Technik, Ressourcenschonung und Energieeffizienz (TREE) (14) (remove)
Document Type
- Conference Object (14) (remove)
Year of publication
- 2018 (14) (remove)
Keywords
- DMA (1)
- DSC (1)
- NEAT (1)
- Neuroevolution (1)
- Surrogate Modeling (1)
- auxiliary power supply (1)
- dimensionality reduction (1)
- draw ratio (1)
- efficiency (1)
- extrusion blow molding (1)
Messkampagnen im Projekt METPVNET zur Verbesserung der PV- Erzeugungsprognose auf Verteilnetzebene
(2018)
Möglichkeiten und Grenzen der Baustoffanalytik und anwendungstechnische Prüfungen an Objekten
(2018)
Untersuchungen zum Einfluss von chemischen Aktivatoren und Templaten auf die Zementhydratation
(2018)
Solar energy plants are one of the key options to serve the rising global energy need with low environmental impact. Aerosols reduce global solar radiation due to absorption and scattering and therewith solar energy yields. Depending on the aerosol composition and size distribution they reduce the direct component of the solar radiation and modify the direction of the diffuse component compared to standard atmospheric conditions without aerosols.
Influence of design of extrusion blow molding (EBM) in terms of extrusion direction set-up and draw ratio as well as process conditions (mold temperature) on storage modulus of high density polyethylene EBM containers was analyzed with dynamic mechanical analysis. All three parameters - mold temperature, flow direction and draw ratio - are statistically significant and lead to relative and absolute evaluation of storage modulus. Furthermore, flow induced changes in crystallinity was analyzed by differential scanning calorimetry. Obtained data on deformation properties can be employed for more sophisticated finite element simulations with the aim to reach more sustainable extrusion blow molding production.
Surrogate-assistance approaches have long been used in computationally expensive domains to improve the data-efficiency of optimization algorithms. Neuroevolution, however, has so far resisted the application of these techniques because it requires the surrogate model to make fitness predictions based on variable topologies, instead of a vector of parameters. Our main insight is that we can sidestep this problem by using kernel-based surrogate models, which require only the definition of a distance measure between individuals. Our second insight is that the well-established Neuroevolution of Augmenting Topologies (NEAT) algorithm provides a computationally efficient distance measure between dissimilar networks in the form of "compatibility distance", initially designed to maintain topological diversity. Combining these two ideas, we introduce a surrogate-assisted neuroevolution algorithm that combines NEAT and a surrogate model built using a compatibility distance kernel. We demonstrate the data-efficiency of this new algorithm on the low dimensional cart-pole swing-up problem, as well as the higher dimensional half-cheetah running task. In both tasks the surrogate-assisted variant achieves the same or better results with several times fewer function evaluations as the original NEAT.