The methods of sensitivity analysis allow to reduce computational cost of multi-iterative optimization procedures by finding the most influential parameters of the particular model. The article presents details of implementation of the numerical library, which is dedicated to sensitivity analysis and can be used by middleware in e-infrastructures. Then, the application of implemented methods to parallel and distributed models is presented on the example of Statistically Similar Representative Volume Element (SSRVE) in the field of metal forming. The influence of parameters, used in the SSRVE methodology, on accuracy of obtained results and performance of calculations is analyzed. The results of sensitivity analysis are presented in the article as well.
CITATION STYLE
Rauch, Łu., Szeliga, D., Bachniak, D., Bzowski, K., & Pietrzyk, M. (2014). Application of sensitivity analysis to grid-based procedure dedicated to creation of SSRVE. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 8500, 364–377. https://doi.org/10.1007/978-3-319-10894-0_26
Mendeley helps you to discover research relevant for your work.