Meu SciELO
Serviços Personalizados
Artigo
Indicadores
- Citado por SciELO
Links relacionados
- Similares em SciELO
Compartilhar
Revista Cubana de Ciencias Informáticas
versão On-line ISSN 2227-1899
Resumo
FERNANDEZ DIAZ, Milenis; ESPINOSA RAMIREZ, José Gabriel e GARCIA-JACAS, César R. Parallel algorithm for spatial interpolation of Ordinary Kriging. Rev cuba cienc informat [online]. 2016, vol.10, n.3, pp. 57-70. ISSN 2227-1899.
ABSTRACT The spatial interpolation methods provide tools for estimating values at unsampled locations using nearby observations. Ordinary Kriging interpolation is one of the most frequently used geostatistical methods for performing spatial analysis. Your goal is to find the Best Linear Unbiased Estimator based on the available data, which are generally insufficient because of the cost of obtaining it. It is characterized by linear algebra expensive operations affecting high execution times and a time complexity of O (MN3). Reducing runtime applications spatial interpolation can be a high priority target, for example, in systems that support rapid decision-making. Various strategies have been applied to reduce high runtimes of Kriging interpolation methods. The techniques of parallel and distributed programming have proven to be a viable alternative for the rapid solution of this type of computational problems. In order to reduce the time associated with spatial interpolation of Ordinary Kriging, was proposed a parallel algorithm based on the use of shared memory programming techniques provided by OpenMP 4.8.2. This algorithm was implemented using C++11 as a programming language and Atlas CLapack as linear algebra library optimized for matrix calculations. The proposed algorithm allows a faster spatial interpolation of Ordinary Kriging, achieving a better utilization of computing resources installed.
Palavras-chave : geostatistical interpolator; Ordinary Kriging; parallel computing; spatial interpolation.