Mathematics of Computation

ISSN 1088-6842(online) ISSN 0025-5718(print)

 

 

Approximations for hand calculators using small integer coefficients


Author: Stephen E. Derenzo
Journal: Math. Comp. 31 (1977), 214-222
MSC: Primary 65D20
MathSciNet review: 0423761
Full-text PDF Free Access

Abstract | References | Similar Articles | Additional Information

Abstract: Methods are presented for deriving approximations containing small integer coefficients. This approach is useful for electronic hand calculators and programmable calculators, where it is important to minimize the number of keystrokes necessary to evaluate the function. For example, the probability $ P(x)$ of exceeding x standard deviations of either sign (Gaussian probability integral) is approximated by

$\displaystyle P(x) \approx {\text{EXP}}\left[ { - \frac{{(83x + 351)x + 562}}{{703/x + 165}}} \right]$

with a relative error less than 0.042

References [Enhancements On Off] (What's this?)

  • [1] 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
  • [2] J. F. HART ET AL., Computer Approximations, Wiley, New York, 1968.
  • [3] For cases where the absolute error $ g(x) - f(x)$ is important, $ w(x)$ is equal to a constant. For cases where the relative error $ (g(x) - f(x))/f(x)$ is important, $ w(x)$ is proportional to $ 1/f(x)$.
  • [4] STEPHEN E. DERENZO, Lawrence Radiation Laboratory Group A Programming Note P-190, Berkeley, Calif., 1969. Most other minimizing codes can also be used.
  • [5] STEPHEN E. DERENZO, Lawrence Berkeley Laboratory Report No. LBL-3804, Berkeley, Calif., March, 1975. (Available from the author.)
  • [6] Henceforth we use the term "best fit coefficients" to mean those that result from the minimization of D (Eq. (2)).
  • [7] The number given after the $ \pm $ symbol is the amount that the corresponding coefficient must be varied from its best fit value to double the value of D, holding all other coefficients at their best fit values.
  • [8] O. KLEIN AND Y. NISHINA, Nature, v. 122, 1928, p. 398. Formulas are also available in the American Institute of Physics Handbook, 3rd ed. (D. E. Gray, Editor), McGraw-Hill, New York, 1972, pp. 8-197.
  • [9] J. H. HUBBELL, Photon Cross Sections, Attenuation Coefficients, and Energy Absorption Coefficients from 10 keV to 100 GeV, Report No. NSRDS-NBS 29, U. S. Nat. Bur. of Standards, 1969.

Similar Articles

Retrieve articles in Mathematics of Computation with MSC: 65D20

Retrieve articles in all journals with MSC: 65D20


Additional Information

DOI: http://dx.doi.org/10.1090/S0025-5718-1977-0423761-X
Article copyright: © Copyright 1977 American Mathematical Society