Dual Algorithms for Pure Network Problems

This paper reports the development of a new algorithmic implementation of the dual algorithm for the capacitated minimum cost network flow problem. Furthermore, it introduces dual reoptimization procedures and compares primal and dual algorithms for the optimization and reoptimization of network problems. The implementation makes use of cut-sets for the efficient execution of the entering variable selection and the selection of the leaving variable is tied to the basis structure at an iteration. Empirical testing of the dual algorithm for optimization shows that it dominates the primal procedure for assignment problems with 400 nodes, transportation problems with more than 600 nodes, and transshipment problems with more than 1500 nodes. Computational testing with typical changes found in decomposition and relaxation procedures indicates the superiority of dual reoptimization over primal reoptimization. For a sequence of parametric changes, as would be typical in large-scale programming techniques, on the average, dual reoptimization is found to require between 75 and 93% fewer pivots and between 20 and 50% less time than primal reoptimization depending on the type of change.

[1]  J. Kennington,et al.  Reoptimization procedures for bounded variable primal simplex network algorithms , 1986 .

[2]  Ellis L. Johnson,et al.  Networks and Basic Solutions , 1966, Oper. Res..

[3]  Jeffery L. Kennington,et al.  Primal simplex network codes: State-of-the-art implementation technology , 1978, Networks.

[4]  J. Orlin.  On the simplex algorithm for networks and generalized networks , 1983 .

[5]  J. Orlin.  Working Paper Alfred P. Sloan School of Management Genuinely Polynominal Simplex and Non-simplex Algorithms for the Minimum Cost Flow Problem Genuinely Polynominal Simplex and Non-simplex Algorithms for the Minimum Cost Flow Problem , 2008 .

[6]  A. Charnes,et al.  Management Models and Industrial Applications of Linear Programming , 1961 .

[7]  Gerald G. Brown,et al.  Design and Implementation of Large-Scale Primal Transshipment Algorithms , 1976 .

[8]  Robert E. Tarjan,et al.  Data structures and network algorithms , 1983, CBMS-NSF regional conference series in applied mathematics.

[9]  Ronald D. Armstrong,et al.  Implementation and analysis of a variant of the dual method for the capacitated transshipment problem , 1980 .

[10]  Fred W. Glover,et al.  Implementation and computational comparisons of primal, dual and primal-dual computer codes for minimum cost network flow problems , 1974, Networks.

[11]  Ronald L. Rardin,et al.  Solving fixed charge network problems with group theory‐based penalties , 1976 .

[12]  A. M. Geoffrion,et al.  Multicommodity Distribution System Design by Benders Decomposition , 1974 .

[13]  R. V. Helgason,et al.  Algorithms for network programming , 1980 .

[14]  R. Rosenthal.  Representing inverses in pure network flow optimization , 1986 .

[15]  Richard V. Helgason,et al.  An Efficient Procedure for Implementing a Dual Simplex Network Flow Algorithm , 1977 .

[16]  Paul A. Jensen,et al.  An advanced dual incremental network algorithm , 1982, Networks.

[17]  Darwin Klingman,et al.  RECENT DEVELOPMENTS IN COMPUTER IMPLEMENTATION TECHNOLOGY FOR NETWORK FLOW ALGORITHMS , 1982 .

[18]  Darwin Klingman,et al.  NETGEN: A Program for Generating Large Scale Capacitated Assignment, Transportation, and Minimum Cost Flow Network Problems , 1974 .