On the round-off error in the method of successive over-relaxation

Author:
M. Stuart Lynn

Journal:
Math. Comp. **18** (1964), 36-49

MSC:
Primary 65.35; Secondary 65.62

DOI:
https://doi.org/10.1090/S0025-5718-1964-0162364-2

MathSciNet review:
0162364

Full-text PDF Free Access

Abstract | References | Similar Articles | Additional Information

Abstract: The asymptotic behavior of the round-off error, which accumulates when the well-known iterative method of (point) successive over-relaxation is used to solve a large-scale system of linear equations, is examined by means of a statistical model. The local round-off errors are treated as independent random variables and expressions for the mean and variance of the accumulated round-off error are obtained, as the number of iterations tends to infinity.

**[1]**A. A. Abramov, ``On the influence of round-off errors in the solution of Laplace's equation,''*Vycis. Matem. i Vycis. Teh.*, v. 1, 1953, p. 37-41. MR**0070267 (16:1156e)****[2]**J. Descloux, ``Note on the round-off error in iterative processes,''*Math. Comp.*, v. 17, 1963, p. 18-27. MR**0152102 (27:2082)****[3]**G. E. Forsythe, ``Note on rounding-off errors,''*SIAM Rev.*, v. 1, 1959, p. 66-67. MR**0099119 (20:5563)****[4]**G. H. Golub, ``The use of Chebyshev matrix polynomials in the iterative solution of linear equations compared to the method of successive over-relaxation,'' Doctoral Thesis, University of Illinois, 1959.**[5]**G. H. Golub,*Bounds for the Round-Off Errors in the Richardson Second-Order Method*, Nordisk Tidskrift for Informations-Behandlung, v. 2, 1962, p. 212-223. MR**0165678 (29:2958)****[6]**G. H. Golub, & J. K. Moore,*ibid*(appendix).**[7]**P. K. Henrici,*Discrete-variable Methods in Ordinary Differential Equations*, John Wiley & Sons, Inc., New York, 1961.**[8]**M. Marcus, ``Basic theorems in matrix theory,''*N.B.S. Appl. Math. Ser.*No. 57, 1960. MR**0109824 (22:709)****[9]**A. Ostrowski, ``On the linear iteration procedures for symmetric matrices,''*Rendic. di Mat. e.d.s. Applicaz*, v. 13, 1954, p. 1-24. MR**0070261 (16:1155e)****[10]**W. Sibagaki, ``On the idea of 'numerical convergence' and its applications,''*Mem. Fac. Sci. Kyushu Univ. Ser. A*, v. 5, 1950, p. 89-97. MR**0039364 (12:537f)****[11]**A. M. Turing, ``Rounding errors in algebraic processes,''*Quart. J. Appl. Math.*, v. 1, 1948, p. 287-307. MR**0028100 (10:405c)****[12]**M. Urabe, ``Convergence of numerical iteration in solution of equations,''*J. Sci. Hiroshima Univ. Ser. A*, v. 19, 1956, p. 479-489. MR**0092225 (19:1081g)****[13]**R. S. Varga,*Matrix Iterative Analysis*, Prentice-Hall, New Jersey, 1962. MR**0158502 (28:1725)****[14]**J. H. Wilkinson, ``Rounding errors in algebraic processes,''*Proc. Int. Conference on Information Processing*, UNESCO, 1959, p. 44. MR**0121976 (22:12703)****[15]**J. H. Wilkinson, Error analysis of direct methods of matrix inversion,''*J. Assoc. Comp. Mach.*, v. 8, 1961, p. 281-330. MR**0176602 (31:874)****[16]**D. M. Young, ``Iterative methods for solving-partial differential equations of the elliptic type,''*Trans. Amer. Math. Soc.*, v. 76, 1954, p. 92-111. MR**0059635 (15:562b)**

Retrieve articles in *Mathematics of Computation*
with MSC:
65.35,
65.62

Retrieve articles in all journals with MSC: 65.35, 65.62

Additional Information

DOI:
https://doi.org/10.1090/S0025-5718-1964-0162364-2

Article copyright:
© Copyright 1964
American Mathematical Society