Journal of applied mathematics & informatics

The Korean Society for Computational and Applied Mathematics (한국전산응용수학회)
권호 표지

A MODIFIED BFGS BUNDLE ALGORITHM BASED ON APPROXIMATE SUBGRADIENTS

  • Guo, Qiang (Business School, University of Shanghai for Science and Technology) ;
  • Liu, Jian-Guo (Research Center of Complex Systems Science, University of Shanghai for Science and Technology)
  • Received : 2009.11.21
  • Accepted : 2009.12.14
  • Published : 2010.09.30
Download PDF
⟨ Previous Next ⟩

Abstract

In this paper, an implementable BFGS bundle algorithm for solving a nonsmooth convex optimization problem is presented. The typical method minimizes an approximate Moreau-Yosida regularization using a BFGS algorithm with inexact function and the approximate gradient values which are generated by a finite inner bundle algorithm. The approximate subgradient of the objective function is used in the algorithm, which can make the algorithm easier to implement. The convergence property of the algorithm is proved under some additional assumptions.

Keywords

References

  1. R. T. Rockafellar, Convex Analysis, Princeton University Press, Princeton, NJ, 1970.
  2. J. B. Hiriart-Urruty, C. Lemarecal, Convex analysis and minimization algorithms, Springer Verlag, Berlin Germany, 1993.
  3. M. Fukushima, L. QI, A globally and superlinearly convergent algorithm for nonsmooth convex minimization, SIAM J. Optim. 6 (1996) 1106-1120. https://doi.org/10.1137/S1052623494278839
  4. R. Mifflin, D. Sun, L. Qi, Quasi-Newton Bundle-Type Methods for Nondifferentiable Convex Optimization, SIAM J. OPTIM 8(2) (1998) 583-603. https://doi.org/10.1137/S1052623496303329
  5. R. Mifflin, A Quasi-Second-Order Proximal Bundle Algorithm, Math. Programming 73 (1996) 51-72.
  6. M. J. D. Powell, On the convergence of the variable metrix algorithm, J. Inst. Math. Appl. 7 (1971), 21-36. https://doi.org/10.1093/imamat/7.1.21
  7. L. C. W. Dixon, Variable metrix algorithms: necessary and sufficient conditions for identical behavior on nonquadratic functions, J. Optim. Theory Appl. 10 (1972) 34-40. https://doi.org/10.1007/BF00934961
  8. R. Byrd, J. Nocedal, Y. Yuan, Global convergence of a class of quasi-Newton methods on convex problems, SIAM J. Numer. Anal. 24 (1987) 1171-1189. https://doi.org/10.1137/0724077
  9. R. H. Byrd, J. Nocedal, A tool for the analysis of quasi-Newton methods with application to unconstrained minimization, SIAM J. Numer. Anal. 26 (1989) 727-739. https://doi.org/10.1137/0726042
  10. D. -H. Li, On the global convergence of DFP method, J. Hunan Univ. (Natural Sciences) 20 (1993) 16-20.
  11. M. J. D. Powell, Some Global Convergence Properties of a Variable Metric Algorithm for Minimization without Exact Line Searches, Nonlinear Programming, SIAM-AMS proceedings Cottle R W and Lemke C E, eds. 9 (1976) 53-72.
  12. J. Werner, Uber die global konvergenze von variablemetric verfahren mit nichtexakter schrittweitenbestimmong, Numer. Math. 31 (1978) 321-334. https://doi.org/10.1007/BF01397884
  13. J. G. Liu, Q. Guo, Global Convergence Properties of the Modified BFGS Method, J. of Appl. Math. & Computing 16 (2004) 195-205. https://doi.org/10.1007/BF02936161
  14. J. G. Liu, Z. Q. Xia, R. D. Ge, and Q. Guo, An modi¯ed BFGS method for non-convex minimization problems (Chinese), Operation Research and Management 13 (2004) 62-65.
  15. Rendong Ge and Zunquan Xia, An ABS Algorithm for Solving Singular Nonlinear System with Rank One Defect, Korea. J. Comp. Appl. Math. 9 (2002) 167-184.
  16. Rendong Ge and Zunquan Xia, An ABS Algorithm for Solving Singular Nonlinear System with Rank Defects, J. of Appl. Math. & Computing 5 (2003) 1-20.
  17. R. Fletcher, An overview of unconstrained optimization, in: E. Spedicato (Ed.), Algorithms for Continous Optimization: The State of the Art, Klumer Acadamic Publisher, Boston, (1994), 109-143.
  18. J. Shen, L. Pang, A quasi-Newton bundle method based on approximate subgradients, J. of Appl. Math. & Computing 23 (2007) 361-367. https://doi.org/10.1007/BF02831983
  19. J. F. Bonnans, J. CH. Gilbert, C. Lemarechal, C. A. Sagastizabal, A family of variable metric proximal methods, Math. Programming 68 (1995) 15-47.
  20. X. Chen, M. Fukushima, Proximal Quasi-Newton Methods for Nondifferentiable Convex Optimization, Applied Mathematics report 95/32, School of Mathematics, The University of New South Wales, Sydney, Australia, 1995.