Probabilistic structural modelling in parallel systems
Abstract eng: Probabilistic methods are used in engineering where a computational model contains random variables. The probabilistic computational methods for reliability calculations can be split e.g. into the three following groups: simulation based techniques, approximation techniques and numerical integration. Series of probabilistic methods were developed; each of them encounters problems of efficiency. The solution may be the use of multiprocessor systems (e.g. supercomputers such as supercomputer Anselm hosted by IT4 Innovations National Supercomputing Center in Ostrava), which enable a large number of computing operations divided by the number of used processing units or cores. Series of probabilistic methods advantageously allows implementing the algorithm of parallelization. Also, a newly developed method Direct Optimized Probabilistic Calculation—“DOProC”, seems for usage in parallel systems as very efficient.
Contributors:
Publisher:
Taylor and Francis Group, London, UK
Conference Title:
Conference Title:
Sixth International Conference on Structural Engineering, Mechanics and Computation
Conference Venue:
Cape Town, South Africa
Conference Dates:
2016-09-05 / 2016-09-07
Rights:
Text je chráněný podle autorského zákona č. 121/2000 Sb.
Record appears in:
Record created 2016-09-20, last modified 2016-09-20
guest ::
