摘要
本文介绍了求线性方程组解的Ｋａｃｚｍａｒｚ法及全并行迭代解法（ＰＧ），它们提出了一种参数辨识和系统建模的快速广义递推估计算法，并给出了收敛性的简单证明。算法的形式简单，计算量少而收敛性能极佳，且便于硬件并行实现。仿真结果也表明了算法的良好性能，可在广阔的领域中得到应用。
In this paper,a kind of iteration method proposed by S.Kaczmarz in 1937 to get the solu-tions of linear equations is reviewed,then we propose a new parallel algorithm(PG).Based on theKaczmarz′s method and PG algorithm,GRE(General Recursive Estimator)is deduced,and its convergency is subsequently proved.The simulation results which show the surprisingly high efficiency and low computing costs of the algorithm illustrate the great power and potential application of it.
出处
《系统仿真学报》
CAS
CSCD
1994年第4期2-8,7,共8页
Journal of System Simulation
关键词
参数估计
广义
递推估计算法
Recursive Identification Predictation Real time
