Abstract
The main result of the paper is the applicability of Dantzig-Wolfe method for Large-Scale Nonlinear Programming with composite (block) structure of the function and constraints. Equivalent transformation of this problem is a task decomposition and coordination. This result allows to propose a new class of decomposition methods, which differ in approximating a feasible solution set of the coordination problem.
The authors propose the modification of the Dantzig-Wolfe algorithm for solving mathematical programming problems, where coordinating solutions is a convex set. It was applied as a decomposition algorithm for a quadratic programming problem. The algorithm was implemented in MS Excel environment and its efficiency was studied and tested. The rate of convergence in the number of global iterations was defined in tests, and it is shown that the proposed algorithm is not significantly different from the Dantzig-Wolfe algorithm in linear block programming.
References
Tsurkov, V.I.: Decomposition in Large-Scale Problems. Monography, Nauka, Moscow (1981)
Mamchenko, O.P., Oskorbin, N.M.: Modeling Hierarchical Systems: Textbook for Higher Education Institutions. Publishing House of ASU, Barnaul (2007)
Oskorbin, N.M., Bogovis, A.V., Zharikov, A.V.: Informational processes coordination of enterprise solutions and computer modeling. Vestn. Novosib. State Univ. Ser. Inf. Technol. 8(1), 54–59 (2010)
Oskorbin, N.M.: On the circuits of block programming. Econ. Math. Methods 5, 964–972 (1981)
Moiseev, N.N., Ivanilov, Y.P., Stolyarova, E.M.: Methods of optimization. Monography, Nauka, Moscow (1978)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2018 Springer International Publishing AG
About this paper
Cite this paper
Oskorbin, N., Khvalynskiy, D. (2018). Decomposition Algorithms for Mathematical Programming and Generalization of the Dantzig-Wolfe Method. In: Silhavy, R., Silhavy, P., Prokopova, Z. (eds) Cybernetics Approaches in Intelligent Systems. CoMeSySo 2017. Advances in Intelligent Systems and Computing, vol 661. Springer, Cham. https://doi.org/10.1007/978-3-319-67618-0_4
Download citation
DOI: https://doi.org/10.1007/978-3-319-67618-0_4
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-67617-3
Online ISBN: 978-3-319-67618-0
eBook Packages: EngineeringEngineering (R0)