论文信息 - Acceleration of vector sequences by multi‐dimensional Δ2 methods

Acceleration of vector sequences by multi‐dimensional Δ2 methods

Nine different algorithms for the acceleration of vector sequences are considered. All have the property that, in one dimension, they reduce to the standard Aitken Δ2 process. Different rules for deciding when to accelerate are also discussed. Theoretical and practical results show a wide range of performance, but it is shown that certain methods can be recommended for use.

A. MacLeod | Allan J. Macleod

[1] A. J. Willmott,et al. Numerical acceleration of thermal regenerator simulations , 1977 .

[2] C. M. Crowe,et al. Convergence promotion in the simulation of chemical processes with recycle-the dominant eigenvalue method , 1971 .

[3] O. C. Zienkiewicz,et al. Accelerated ‘relaxation’ or direct solution? Future prospects for fem , 1985 .

[4] P. Wynn,et al. Acceleration techniques for iterated vector and matrix problems : (mathematics of computation, _1_6(1962), nr 79, p 301-322) , 1962 .

[5] R. Pennacchi,et al. Le trasformazioni razionali di una successione , 1968 .

[6] Donald G. M. Anderson. Iterative Procedures for Nonlinear Integral Equations , 1965, JACM.

[7] Bruce M. Irons,et al. A version of the Aitken accelerator for computer iteration , 1969 .

[8] A. Jennings. Accelerating the Convergence of Matrix Iterative Processes , 1971 .