The Parallelization Research And Implementation Of GIS Spacial Site Selection Algorithms

The spatial site selection is widely used in daily life, and comparing to the traditional site selection method, using Geographic Information System(GIS) technology for site selection shows absolute advantage and gains a wide range of popularization. However,due to the increasingly mass spatial data and complicated site selection model, the current GIS site selection method appears limitation and insufficiency in storage capacity and processing speed. In order to improve the efficiency of GIS site selection, this paper presents a parallel GIS spatial site selection project, which aims at efficiently parallel the vector buffer algorithm and vector overlay algorithm, both being basal and crucial in the GIS site selection analysis, as a result, meeting the demand of efficient and fast site selection analysis.Firstly, in the multi-core processors and cluster environment, the paper uses Message Passing Interface(MPI) and Single Program Multiple Data(SPMD) to realize the parallel by exchanging information and coordinating the pace to control execution.Secondly, based on the existing vector spatial data partitioning method, this paper proposes two efficient partition strategies, respectively being the arc partition for the buffer algorithm and horizontal stripe partition for the overlay algorithm, and then designs parallel buffer and parallel overlay algorithm, based on the two partition strategy.Thirdly, doing the test and analysis of parallel algorithm, with the help of the open source GIS software, namely Geographic Resources Analysis Support System(GRASS),by adding the parallel interface to the GRASS loosely coupled modular system structure,and building the parallel algorithms library, which is equivalent with the original serial algorithm.Finally, based on the designed parallel buffer analysis algorithm and parallel overlay algorithm, using the park site selection as a simulation application case to verify the method’s correctness and effectiveness. The case proved that the method has a good application prospect.