A Data Decomposition Based Load Balance Method in Parallel Map Overlay Analysis
Ji Mi · Geomatics & Spatial Information Technology · 2015
Map overlay analysis is a computing intensive algorithm and parallel computing is a feasible way that could accelerate the running speed. The paper studies the load balance method in parallel map overlay analysis. First,with respect to the condition of IO competition,a load balance index is defined. Then based on the divide and conquer approach,geographic data is distributed on the parallel computing system. The Hilbert spatial index is applied in detailed data decomposition to the child node in parallel map overlay method,that makes data parallel possible. The results indicates that using the spatial index data distribution method it is fast to decide the subset of map data in parallel computing system,and computing tasks can explicitly be allocated to each child node.