A Statistical Review of the Sandgren-Ragsdell Comparative Study

  • Eric Sandgren
Conference paper
Part of the Lecture Notes in Economics and Mathematical Systems book series (LNE, volume 199)


A statistical analysis of the solution times of the algorithms in the Sandgren-Ragsdell study is conducted. An analysis of variance is performed to demonstrate that there is statistical evidence that selected codes are superior to others on the basis of their relative solution times. A logarithmic transformation is used to produce a seminormal distribution of the solution times with a variance assumed to be equal for all of the algorithms. A paired comparison is then conducted on the differences in the mean logarithmic solution times for each of the algorithms over the entire test problem set. The selected confidence level for all comparisons was fixed at 95%. The factors contributing to the success of this analysis are discussed as well as the additional data which would be required to conduct this type of analysis in general.


Test Problem Solution Time Problem Population Linear Programming Algorithm Corporation Information System 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Lee, H. K., and Ravindran, A., “A Comparison of Five Algorithms for Solving Quadratic Programming Problems,” Working Paper, School of Industrial Engineering, Purdue University, August 1975.Google Scholar
  2. 2.
    Kuhn, H. W., and Quandt, R. E., “An Experimental Study of the Simplex Method,” Proc. Sympos. Appl. Math. 15, 1963, pp. 107–124.Google Scholar
  3. 3.
    Sandgren, E., “The Utility of Nonlinear Programming Algorithms,” Ph.D. Dissertation, Purdue University, December 1977.Google Scholar
  4. 4.
    Colville, A. R., “A Comparative Study of Nonlinear Programming Codes,” Technical Report No. 320–2949, IBM New York Scientific Center, June 1968.Google Scholar
  5. 5.
    Eason, E. D., and Fenton, R. G., “Testing and Evaluation of Numerical Methods for Design Optimization,” UTME-TP 7204, University of Toronto, September 1972.Google Scholar
  6. 6.
    Fattler, J. E., Sin, Y. T., Root, R. R., Ragsdell, K. M., and Reklaitis, G. V., “On the Computational Utility of Polynomial Geometric Programming Solution Methods,” Presented at the Tenth International Symposium on Mathematical Programming, Montreal, August 1979.Google Scholar
  7. 7.
    Sandgren, E., and Ragsdell, K. M., “The Utility of Nonlinear Programming Algorithms: A Comparative Study- Parts I and II,” ASME Journal of Mechanical Design, Vol. 102, No. 3, July 1980, pp. 540–551.CrossRefGoogle Scholar
  8. 8.
    Box, E. P., Hunter, W. G., and Hunter, J. S., Statistics for Experiments, Wiley, New York, 1978.Google Scholar
  9. 9.
    Gibra, I. N., Probability and Statistical Inference for Scientists and Engineers, Prentice-Hall, N.J., 1973.Google Scholar
  10. 10.
    Tukey, J.W., “Comparing Individual Means in the Analysis of Variance,” Biometrics, 5. 99, 1949.CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1982

Authors and Affiliations

  • Eric Sandgren
    • 1
  1. 1.Information Systems DivisionIBM CorporationLexingtonUSA

Personalised recommendations