Low-order approximations for the normal probability integral and the error function

Carta, David G.

doi:10.1090/S0025-5718-1975-0368389-3

Low-order approximations for the normal probability integral and the error function
HTML articles powered by AMS MathViewer

by David G. Carta PDF

Math. Comp. 29 (1975), 856-862 Request permission

Abstract:

Rational fractions of the form $0.5/{(a + bx + \ldots )^{2q}}$ are used to evaluate the function of interest. Polynomials of from third to sixth order are derived which achieve absolute errors ranging from 0.01 to 0.000001 for all (real) positive x, and relative errors of from 0.1 to 0.00001 for (real) positive x less than 3.1, 4.0, and 5.2. Denominator coefficients are calculated by linearizing the rational fraction about progressively improved nominal solutions and using linear programming to solve the resulting linear minimax problems.

References

Record of Proc., The Seventh Annual Simulation Symposium

Cecil Hastings Jr., Approximations for digital computers, Princeton University Press, Princeton, N. J., 1955. Assisted by Jeanne T. Hayward and James P. Wong, Jr. MR 0068915

Computer Approximations

Philip Rabinowitz, Applications of linear programming to numerical analysis, SIAM Rev. 10 (1968), 121–159. MR 226810, DOI 10.1137/1010029
Survey of numerical analysis, McGraw-Hill Book Co., Inc., New York-Toronto-London, 1962. MR 0135221
È. G. Belaga, Some problems involved in the calculation of polynomials, Dokl. Akad. Nauk SSSR 123 (1958), 775–777 (Russian). MR 0105192
L. A. Lyusternik, O. A. Chervonenkis, and A. R. Yanpol′skii, Handbook for computing elementary functions, Pergamon Press, Oxford-Edinburgh-New York, 1965. Translated from the Russian by G. J. Tee; Translation edited by K. L. Stewart. MR 0183102
Donald E. Knuth, The art of computer programming. Vol. 2: Seminumerical algorithms, Addison-Wesley Publishing Co., Reading, Mass.-London-Don Mills, Ont., 1969. MR 0286318