Computer Engineering and Applications ›› 2011, Vol. 47 ›› Issue (2): 49-51.DOI: 10.3778/j.issn.1002-8331.2011.02.016

• 研究、探讨 • Previous Articles     Next Articles

Parallel algorithm for solving nonlinear system of equations

WANG Bao,SUN Qin   

  1. School of Aeronautics,Northwestern Polytechnical University,Xi’an 710072,China
  • Received:2009-05-08 Revised:2009-09-07 Online:2011-01-11 Published:2011-01-11
  • Contact: WANG Bao


汪 保,孙 秦   

  1. 西北工业大学 航空学院,西安 710072
  • 通讯作者: 汪 保

Abstract: A parallel algorithm for solving nonlinear system of equations on distributed-memory multi-computers is presented.To parallelly solve nonlinear system of equations,Jacobi matrix in Newton method is properly splitted.In theory,analysis of convergence about this algorithm is given.Some numerical results on HP rx2600 cluster show that the algorithm’s parallel efficiency exceeds 70%.

Key words: nonlinear system of equations, parallel algorithm, HP rx2600 cluster

摘要: 提出了一种在分布式环境下求解非线性方程组的并行算法,该算法将Newton迭代法中的Jacobi矩阵进行适当的分裂,使得Newton迭代法具有很好的并行性。并在理论上进行了收敛性分析。在HP rx2600集群上进行的数值实验结果表明并行效率达70%以上。

关键词: 非线性方程组, 并行算法, HP rx2600集群

CLC Number: