Abstract
The limited memory BFGS method (L-BFGS) is an adaptation of the BFGS method for large-scale unconstrained optimization. However, The L-BFGS method need not converge for nonconvex objective functions and it is inefficient on highly ill-conditioned problems. In this paper, we proposed a regularization strategy on the L-BFGS method, where the used regularization parameter may play a compensation role in some sense when the condition number of Hessian approximation tends to become ill-conditioned. Then we proposed a regularized L-BFGS method and established its global convergence even when the objective function is nonconvex. Numerical results show that the proposed method is efficient.
Similar content being viewed by others
References
Al-Baali, M.: Improved Hessian approximations for limited memory BFGS method. Numer. Algoritm. 22, 99–112 (1999)
Andrei, N.: An unconstrained optimization test functions collection. Adv. Model. Optim. 10, 147–161 (2008)
Byrd, R.H., Nocedal, J., Yuan, Y.X.: Global convergence of a class of quasi-Newton methods on convex problems. SIAM J. Numer. Anal. 24, 1171–1190 (1987)
Byrd, R.H., Nocedal, J.: A tool for the analysis of quasi-Newton methods with application to unconstrained minimization. SIAM J. Numer. Anal. 26, 727–739 (1989)
Chen, X.J.: Convergence of the BFGS method for \(LC^{1}\) convex constrained optimization. SIAM J. Control Optim. 34, 2051–2063 (1996)
Dai, Y.H.: Convergence properties of the BFGS algorithm. SIAM J. Optim. 13, 693–701 (2002)
Dennis, J.E., Moré, J.J.: A characterization of superlinear convergence and its application to quasi-Newton methods. Math. Comput. 28, 549–560 (1974)
Dennis, J.E., Moré, J.J.: Quasi-Newton methods, motivation and theory. SIAM Rev. 19, 46–89 (1977)
Dolan, E.D., Moré, J.J.: Benchmarking optimization software with performance profiles. Math. Program. 91, 201–213 (2002)
Fletcher, R.: Practical Methods of Optimization, 2nd edn. Wiley, Chichester (1987)
Li, D.H., Fukushima, M.: On the global convergence of the BFGS method for nonconvex unconstrained problems. SIAM J. Optim. 11, 1054–1064 (2001a)
Li, D.H., Fukushima, M.: A modified BFGS method and its global convergence in nonconvex minimization. J. Comput. Appl. Math. 129, 15–35 (2001b)
Liu, D.C., Nocedal, J.: On the limited memory BFGS method for large-sclae optimization. Math. Program. 45, 503–528 (1989)
Liu, T.W., Li, D.H.: A pratical update criterion for SQP method. Optim. Methods Soft. 2, 253–266 (2007)
Mascarenhas, W.F.: The BFGS method with exact line searches fails for nonconvex objective functions. Math. Program. A 99, 49–61 (2004)
moré, J.J., Garbow, B.S., Hillstrome, K.E.: Testing unconstrained optimization software. ACM Trans. Math. Soft. 7, 17–41 (1981)
Morales, J.L.: A numerical study of limited memory BFGS methods. Appl. Math. Lett. 15, 481–487 (2002)
Nash, S.H., Nocedal, J.: A numerical study of the limited memory BFGS method and the truncated-Newton method for large-scale optimization. SIAM J. Optim. 1, 358–372 (1991)
Nocedal, J.: Updating quasi-Newton matrices with limited storge. Math. Comput. 35, 773–782 (1980)
Nocedal, J.: Theory of algorithms for unconstrained optimization. Acta Numer. 1, 199–242 (1992)
Nocedal, J., St. Wright, J.: Numerical Optimization. Springer, New York (2000)
Powell, M.J.D.: Quadratic termination properties of minimization algorithm, Part I and Part II . J. Inst. Math. Appl. 10, 333–357 (1972)
Powell, M.J.D.: Some convergence properties of a variable mertric algorithm for minimization without exact line search. In: Cottle, R.W., Lemke, C.E. (eds.) Nonlinear Programming, SIAM-AMS Proceedings, vol. IX, pp 53–72. SIAM, Philadelphia (1976)
Powell, M.J.D.: On the convergence of the DEP algorithm for unconstrained optimization when there are only two variables. Math. Program. B 87, 281–301 (2000)
Pu, D., Yu, W.: On the convergence of the DFP algorithm. Ann. Oper. Res. 24, 175–184 (1990)
Wei, Z.X., Yu, G.H., Yuan, G.L., et al: The superlinear convergence of a modified BFGS-type method for unconstrained optimization. J. Comput. Optim. Appl. 29, 315–332 (2004)
Xiao, Y.H., Wei, Z.X., Wang, Z.G.: A limited memory BFGS-type method for large-scale unconstrained optimization. Comput. Math. Appl. 56, 1001–1009 (2008)
Yu, G.H., Wei, Z.X., Guan, L.T.: A modified BFGS-trust region method for nonconvex minimization. Pac. J. Optim. 2, 119–133 (2006)
Yang, Y.T., Xu, C.X.: A compact limited memory method for large-sclae unconstrained optimization. Eur. Oper. Res. 180, 48–56 (2007)
Yuan, Y.: Numerical Methods for Nonlinear Programming. Shanghai Scientific and Technical Publishers, Shanghai (1993). In Chinese
Zhou, W., Zhang, L.: Global convergence of a regularized factorized quasi-Newton method for nonlinear least squares problems. Comput. Appl. Math. 29, 195–214 (2010)
Author information
Authors and Affiliations
Corresponding author
Additional information
This work was partially supported by the National Natural Sciences Foundation of China (No.11271115) and the Hunan Natural Sciences Foundation of China (No.13JJ3040)
Rights and permissions
About this article
Cite this article
Liu, TW. A regularized limited memory BFGS method for nonconvex unconstrained minimization. Numer Algor 65, 305–323 (2014). https://doi.org/10.1007/s11075-013-9706-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11075-013-9706-y