Abstract
We use quadratic penalty functions along with some recent ideas from linearl 1 estimation to arrive at a new characterization of primal optimal solutions in linear programs. The algorithmic implications of this analysis are studied, and a new, finite penalty algorithm for linear programming is designed. Preliminary computational results are presented.
Similar content being viewed by others
References
Bartels RH (1980) A penalty linear programming method using reduced-gradient basis-exchange techniques.Linear Algebra Appl 29:17–32
Bertsekas DP (1975) Necessary and sufficient conditions for a penalty method to be exact.Mathematical Programming 9:87–99
Bixby RE (1990) Implementing the simplex method: Part I: Introduction and Part II: The initial basis. Technical Report TR90-32, Rice University, (also in ORSA J on Computing)
Chebotarev SP (1973) Variation of the penalty coefficient in linear programming problems.Automation and Control 7:102–107
Conn AR (1976) Linear programming via a nondifferentiable penalty function. SIAM J Numer Anal 13:145–154
Eaton JW (1993) OCTAVE: A high-level interactive language for numerical computations. Manuscript University of Texas at Austin
Gill PE, Hammarling S, Murray W, Saunders MA, Wright MH (1986) User's guide for LSSOL (Version 1.0): A fortran package for constrained linear least squares and convex quadratic programming. Report SOL 86-1, Department of Operations Research, Stanford University, Stanford, CA
Huber P (1981) Roubust statistics. John Wiley, New York
Madsen K, Nielsen HB (1990) Finite algorithms for robust linear regression. BIT 30:682–699
Madsen K, Nielsen HB, Pinar MC (1993) New characterizations ofl 1 solutions of overdetermined linear systems. Operations Research Letters 16:159–166
Mangasarian OL (1986) Some applications of penalty functions in mathematical programming. Lecture Notes in Mathematics, No 1190. In: Conti R, De Giorgi E, Giannessi F (eds), Optimization and Related Fields. Springer-Verlag, Berlin 307–329
Murty KG (1976) Linear and combinatorial programming. John Wiley & Sons, New York
Nielsen HB (1990) AAFAC: A package of fortran 77 subprograms for solvingA T Ax = c. Report NI-90-11. Institute for Numerical Analysis, Technical University of Denmark
Pietrzykowski T (1961) Application of the steepest ascent method to linear programming. PRACE ZAMA 11
Pinar MÇ Piecewise-linear pathways to the optimal set in linear programming. Technical Report, Bilkent University, 1996
Author information
Authors and Affiliations
Additional information
Research supported by grant No. 11-0505 from the Danish Natural Science Research Council SNF.
Rights and permissions
About this article
Cite this article
Pinar, M.Ç. Linear programming via a quadratic penalty function. Mathematical Methods of Operations Research 44, 345–370 (1996). https://doi.org/10.1007/BF01193936
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF01193936