TY - JOUR
T1 - A theoretical approach to the use of cyberinfrastructure in geographical analysis
AU - Wang, Shaowen
AU - Armstrong, Marc P.
N1 - Funding Information:
This research was supported in part by the National Science Foundation through the TeraGrid computation resource award: TG-SES070007N. We are grateful for the insightful comments of the editor and three anonymous reviewers.
PY - 2009
Y1 - 2009
N2 - This paper presents a theoretical approach that has been developed to capture the computational intensity and computing resource requirements of geographical data and analysis methods. These requirements are then transformed into a common framework, a grid-based representation of a spatial computational domain, which supports the efficient use of emerging cyberinfrastructure environments. Two key types of transformational functions (data-centric and operation-centric) are identified and their relationships are explained. The application of the approach is illustrated using two geographical analysis methods: inverse distance weighted interpolation and the G* i(d) spatial statistic. We describe the underpinnings of these two methods, present their conventional sequential algorithms, and then address their latent parallelism based on a spatial computational domain representation. Through the application of this theoretical approach, the development of domain decomposition methods is decoupled from specific high-performance computer architectures and task scheduling implementations, which makes the design of generic parallel processing solutions feasible for geographical analyses.
AB - This paper presents a theoretical approach that has been developed to capture the computational intensity and computing resource requirements of geographical data and analysis methods. These requirements are then transformed into a common framework, a grid-based representation of a spatial computational domain, which supports the efficient use of emerging cyberinfrastructure environments. Two key types of transformational functions (data-centric and operation-centric) are identified and their relationships are explained. The application of the approach is illustrated using two geographical analysis methods: inverse distance weighted interpolation and the G* i(d) spatial statistic. We describe the underpinnings of these two methods, present their conventional sequential algorithms, and then address their latent parallelism based on a spatial computational domain representation. Through the application of this theoretical approach, the development of domain decomposition methods is decoupled from specific high-performance computer architectures and task scheduling implementations, which makes the design of generic parallel processing solutions feasible for geographical analyses.
KW - Computational transformation
KW - Cyberin frastructure
KW - Geographical analysis
KW - Parallel processing
KW - Spatial computational domain
UR - http://www.scopus.com/inward/record.url?scp=62149136702&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=62149136702&partnerID=8YFLogxK
U2 - 10.1080/13658810801918509
DO - 10.1080/13658810801918509
M3 - Article
AN - SCOPUS:62149136702
SN - 1365-8816
VL - 23
SP - 169
EP - 193
JO - International Journal of Geographical Information Science
JF - International Journal of Geographical Information Science
IS - 2
ER -