AnimplementationofKarmarkarsalgorithmforlinearprogramming.pdf

Mathematical Programming 44 (1989) 297-335 297 North-Holland AN IMPLEMENTATION OF KARMARKARS ALGORITHM FOR LINEAR PROGRAMMING Ilan A D L E R , Mauricio G.C. R E S E N D E * and G e r a l d o V E I G A Department of Industrial Engineering and Operations Research, University o[ California, Berkeley, CA 94720, USA N a r e n d r a K A R M A R K A R A TT Bell Laboratories, Murray Hill, NJ 07974, USA Received 29 June 1987 Revised manuscript received 14 March 1989 This paper describes the implementation of power series dual affine scaling variants of Karmarkars algorithm for linear programming. Based on a continuous version of Karmarkars algorithm, two variants resulting from first and second order approximations of the continuous trajectory are implemented and tested. Linear programs are expressed in an inequality form, which allows for the inexact computation of the algorithms direction of improvement, resulting in a significant computational advantage. Implementation issues particular to this family of algorithms, such as treatment of dense columns, are discussed. The code is tested on several standard linear program- ming problems and compares favorably with the simplex code MINOS4.O. Key words: Linear programming, Karmarkars algorithm, interior point methods. I. Introduction We describe in this p a p e r a family of interior point p o w e r series affine scaling algorithms based on the linear p r o g r a m m i n g algorithm presented by K a r m a r k a r (1984). Two algorithms from this family, corresponding to first and