Modification of iterative processes for improved convergence characteristics
暂无分享,去创建一个
A method for improving the convergence rate of iterative processes is presented. This method monitors the sequence of approximations produced by an iterative process, and generates a new iterative process, or sequence of processes, each with more rapid convergence characteristics than the previous one. This method of process modification can be applied to many iterative techniques currently in use. Examples of application of the method are given, using both nonlinear and linear systems. In one of these examples it is shown that the method can be applied to a divergent iterative process, producing a convergent one. It is also shown how the technique can be applied to equations of one variable, yielding a method with convergence characteristics similar to those of Newton's method, but without the explicit calculation of derivatives.
[1] Alan Jennings,et al. Matrix Computation for Engineers and Scientists , 1977 .
[2] O. C. Zienkiewicz,et al. Note on the ‘Alpha’‐constant stiffness method for the analysis of non‐linear problems , 1972 .
[3] J. H. Wilkinson. The algebraic eigenvalue problem , 1966 .
[4] A. P. Kabaila,et al. Finite element post‐buckling analysis for frames , 1967 .