Abstract
By attacking the linear programming problems from their dual side, a new general algorithm for linear programming is developed. At each iteration, the algorithm finds a feasible descent search direction by handling a least square problem associated with the dual system, using QR decomposition technique. The new method is a combination of pivot method and interior-point method. It in fact not only reduces the possibility of difficulty arising from degeneracy, but also has the same advantages as pivot method in warm-start to resolve linear programming problems. Numerical results of a group of randomly constructed problems are very encouraging.
Similar content being viewed by others
References
Forrest J J H, Goldfarb D. Steepest-edge simplex algorithms for linear programming, Mathematical Programming, 1992, 57:341–374.
Tibor Illés, Tamás Terlaky. Pivot versus interior points methods: Pros and cons, European Journal of Operational Research, 2002, 140:170–190.
Pan P Q. A projective simplex method for linear programming, Linear Algebra and Its Applications, 1999, 29(2):99–125.
Pan P Q, Li Wei, Wang Yang. A phase-I algorithm using the most-obtuse-angle rule for the basis-deficiency-allowing dual simplex method, OR Transaction, 2004, 8(3):88–96.
Pan P Q, Li Wei. A non-monotone phase-1 method in linear programming, Journal of Southeast University, 2003, 19(3):293–296.
Li Wei. A new simplex-like algorithm for linear programming, Mathematical Theory and Applications, 2003, 23(3):118–122.
Li Wei. A note on two direct methods in linear programming, European Journal of Operational Research, 2004, 158(1):262–265.
Anderson, E D, Ye Yinyu. Combining interior-point and pivoting algorithms for linear programming, Management Science, 1999, 42(12):1719–1731.
Author information
Authors and Affiliations
Additional information
Supported by the National Natural Science Foundation of China (10371028), Science Foundation of Zhejiang Bureau of Education (20030622) and Science Foundation of Hangzhou University of Electronic Technology (KYS091504025).
Rights and permissions
About this article
Cite this article
Wei, L., Guangting, C. A new least square algorithm for linear programming. Appl. Math.- J. Chin. Univ. 21, 214–222 (2006). https://doi.org/10.1007/BF02791347
Received:
Issue Date:
DOI: https://doi.org/10.1007/BF02791347