Abstract
In this paper, the alternative duality schemes for mathematical programming problems are considered. These schemes are based on the Lagrange function regularized by Tikhonov in primal and dual variables simultaneously. Earlier such schemes were investigated for convex programming problems only, and the conditions were found which guarantee a convergence of both primal and dual components of the saddle point of the regularized Lagrange function to the optimal sets of primal and dual problems respectively. In addition, some accuracy estimates were obtained for deviation of only one of these components from the normal solution (solution with minimal the Euclidean norm) of the corresponding problem, primal or dual. Unfortunately, these estimates are valid only if primal and dual parameters of regularization have a different order of smallness. In this article, the linear case is investigated in detail. It is shown that for linear programming problem both mentioned sequences converge to the normal solutions of primal and dual programs simultaneously, and it is not essential for such a convergence whether the regularization parameters have a different or the same order of smallness. Also, the alternative accuracy estimates are presented which appear to be more precise and efficient in comparison with the estimates known for a convex case.
This work was supported by Russian Science Foundation, grant N 14–11–00109.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
It follows from well-known Hoffman’s lemma [11] estimating the distance of an arbitrary point to the polyhedral set defined by the system of linear inequalities (and the optimal set of a linear programming problem belongs to such a class) in terms of deviations of this point from each inequality separately.
References
Tikhonov, A.N., Arsenin, V.Ya.: Methods for Solutions of Ill-Posed Problems. Nauka, Moscow (1979). Wiley, New York (1981)
Vasil’ev, F.P.: Optimization Methods, vols. 1, 2. MTsNMO, Moscow (2011). In Russian
Rockafellar, R.T.: Augmented Lagrange multiplier functions and duality in nonconvex programming. SIAM J. Control 12(2), 268–285 (1974)
Evtushenko, Ju.: Numerical Optimization Technique. Optimization Software, Inc. Publishing Division, New York (1985)
Gol’shtein, E.G., Tret’yakov, N.V.: Modified Lagrange Functions. Theory and Methods of Optimization. Nauka, Moscow (1989). In Rusian
Roos, C., Terlaky, T.: Theory and Algorithms for Linear Optimization: An Interior Point Approach. Wiley, New York (1997)
Eremin, I.I., Mazurov, Vl.D., Astaf’ev, N.N.: Improper Problems of Linear and Convex Programming. Nauka, Moscow (1983). In Russian
Skarin, V.D.: Regularization of the min-max problems occurring in convex programming. USSR Comput. Math. Math. Phys. 17(6), 65–78 (1977). https://doi.org/10.1016/0041-5553(77)90173-2
Skarin, V.D.: On method of regularization for infeasible problems of convex programming. Izv. VUZov. Mathematica 12, 81–88 (1995). Russian Math. (Iz. VUZ), 39(12), 78–85 (1995)
Eremin, I.I.: Theory of Linear Optimization. Inverse and Ill-Posed Problems Series. VSP, Utrecht, Boston, Koln, Tokyo (2002)
Hoffman, A.J.: On approximate solutions of systems of linear inequalities. J. Res. Nat. Bur. Stand. 49, 263–265 (1952)
Guddat, J.: Stability in convex quadratic programming. Mathematische Operationsforschung und Statistik 8, 223–245 (1976)
Dorn, W.S.: Duality in quadratic programming. Q. Appl. Math. 18, 407–413 (1960)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2018 Springer International Publishing AG, part of Springer Nature
About this paper
Cite this paper
Popov, L.D. (2018). On Accuracy Estimates for One Regularization Method in Linear Programming. In: Eremeev, A., Khachay, M., Kochetov, Y., Pardalos, P. (eds) Optimization Problems and Their Applications. OPTA 2018. Communications in Computer and Information Science, vol 871. Springer, Cham. https://doi.org/10.1007/978-3-319-93800-4_14
Download citation
DOI: https://doi.org/10.1007/978-3-319-93800-4_14
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-93799-1
Online ISBN: 978-3-319-93800-4
eBook Packages: Computer ScienceComputer Science (R0)