基于MapReduce的并行石漠化CA模型

摘要

针对石漠化演化模拟预测CA模型在单机上训练和运行时间较长的问题。给出了MapReduce编程模型实现的并行化石漠化CA模型，并在用普通PC搭建的Hadoop集群上进行研究实验。实验结果表明，在Hadoop集群上实现的MapReduce并行化石漠化CA模型具有较好的加速比。%According to the problem that the CA model of rocky desertification evolution simulation and prediction runs in a sin-gle PC has a long time. This paper gives a parallel rocky desertification CA model based on MapReduce programming model, and performs research test in a Hadoop cluster built by the ordinary PC, the experimental results show that the parallel MapRe-duce rocky desertification CA model implemented in the Hadoop group has better speedup.