Skip to main content
Log in

Metric-Based Symmetric Rank-One Updates

  • Published:
Computational Optimization and Applications Aims and scope Submit manuscript

Abstract

Metric-based SR1 updates which are stabilized by a variationalrelaxation of the quasi-Newton relation are examined. Thisinvestigation reveals an interesting and surprising connection to theorigin of quasi-Newton methods as first formulated by Davidon [1]. Anextended version of Davidon's original direct prediction SR1 updateis shown to be self-complementary and to possess a finite terminationproperty on quadratics, and limited-memory versions of the update areshown to be globally convergent. Variants of this update are testednumerically and compared to several other metric-based SR1 variantsand the BFGS update. Finally, metric-based stabilizations of the SR1update are critiqued in general, and a promising new model-basedstrategy recently developed is briefly described.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

References

  1. H.L. Anderson and W.C. Davidon, "Machine analysis of pion scattering by the maximum likelihood method," Il Nuovo Cimento, Serie X, vol. 5, pp. 1238–1255, 1957.

    Google Scholar 

  2. R.K. Brayton and J. Cullum, "An algorithm for minimizing a differentiable function subject to box constraints and errors," J.O.T.A., vol. 29, pp. 521–558, 1979.

    Google Scholar 

  3. C.G. Broyden, "Quasi-Newton methods and their application to function minimization," Mathematics of Computation, pp. 368–381, 1967.

  4. R.H. Byrd, "A multiobjective characterization of the Broyden class," presented at ORSA/TIMS, New York, 1990.

  5. A.R. Conn, N.I.M. Gould, and Ph.L. Toint, "Convergence of quasi-Newton matrices generated by the symmetric rank one update," Mathematical Programing, vol. 50, pp. 177–195, 1991.

    Google Scholar 

  6. W.C. Davidon, "Variable metric methods for minimization," Research and Development Report ANL-5990 (Rev.), Argonne National Laboratory, Argonne, Illinois, 1959. (Reprinted in SIAM J. Optimization, vol. 1, 1991.)

    Google Scholar 

  7. W.C. Davidon, "Variance algorithm for minimization," Computer Journal, vol. 10, pp. 406–410, 1968.

    Google Scholar 

  8. W.C. Davidon, "Optimally conditioned optimization algorithms without line searches," Mathematical Programming, vol. 9, pp. 1–30, 1975.

    Google Scholar 

  9. W.C. Davidon, R.B. Mifflin, and J.L. Nazareth, "Some comments on notation for quasi-Newton methods," Optima, vol. 32, pp. 3–4, 1991.

    Google Scholar 

  10. J.E. Dennis and H. Wolkowicz, "Sizing and least change secant updates," CORR Report 90-02, Department of Combinatorics and Optimization, University of Waterloo, Ontario, Canada, 1990.

    Google Scholar 

  11. A.V. Fiacco and G.P. McCormick, Nonlinear Programming: Sequential Unconstrained Minimization Techniques, John Wiley: N.Y., 1968.

    Google Scholar 

  12. H. Khalfan, R.H. Byrd, and R.B. Schnabel, "A theoretical and experimental study of the symmetric rank one update," SIAM J. Optimization, vol. 3, pp. 1–24, 1993.

    Google Scholar 

  13. R.B. Mifflin and J.L. Nazareth, "The least prior deviation quasi-Newton update," Mathematical Programming, vol. 65, pp. 247–261, 1994.

    Google Scholar 

  14. J.J. Moré, B.S. Garbow, and K.E. Hillstrom, "Testing unconstrained optimization software," ACM Trans. on Mathematical Software, vol. 7, pp. 17–41, 1981.

    Google Scholar 

  15. J.J. Moré and D.J. Thuente, "On linesearch algorithms with guaranteed sufficient decrease," MCS-P153-0590, Argonne National Laboratory, Argonne, IL, 1990.

    Google Scholar 

  16. J.L. Nazareth, "An alternative variational principle for variable metric updating," Mathematical Programming, vol. 30, pp. 99–104, 1984.

    Google Scholar 

  17. J.L. Nazareth, "The Newton and Cauchy perspectives on computational nonlinear optimization," SIAM Review, vol. 36, pp. 215–225, 1994.

    Google Scholar 

  18. J. Nocedal, "Theory of algorithms for unconstrained optimization," Acta Numerica, vol. 1, pp. 199–242, 1991.

    Google Scholar 

  19. M.R. Osborne and L.P. Sun, "A new approach to the Symmetric-Rank-One updating algorithm," Report NMO/01, Department of Statistics, Australian National University, Canberra, Australia, 1988.

    Google Scholar 

  20. J.D. Pearson, "Variable metric methods of minimization," Computer Journal, vol. 12, pp. 171–178, 1969.

    Google Scholar 

  21. B. Smith, "Metric and model based line search symmetric rank one updates," Ph.D. Thesis, Department of Pure and Applied Mathematics, Washington State University, 1995.

    Google Scholar 

  22. P. Wolfe, "Convergence conditions for ascent methods," SIAM Review, vol. 11, pp. 226–2235, 1969.

    Google Scholar 

  23. P. Wolfe, "Convergence conditions for ascent methods II: Some corrections,"SIAM Review, vol. 13, pp. 185–188, 1971.

    Google Scholar 

  24. H. Wolkowicz, "Measures for symmetric rank-one updates," Mathematics of Operations Research, vol. 19, pp. 815–830, 1994.

    Google Scholar 

  25. H. Wolkowicz and Q. Zhao, "An all-inclusive efficient region of updates for least change secant methods," SIAM J. Optimization, vol. 5, pp. 171–191, 1995.

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Rights and permissions

Reprints and permissions

About this article

Cite this article

Smith, B., Nazareth, J. Metric-Based Symmetric Rank-One Updates. Computational Optimization and Applications 8, 219–244 (1997). https://doi.org/10.1023/A:1008645321602

Download citation

  • Issue Date:

  • DOI: https://doi.org/10.1023/A:1008645321602

Navigation