Abstract
Decomposition algorithms for block-angular linear programs give rise to a natural, coarse-grained parallelism that can be exploited by processing the subproblems concurrently within a distributed-memory environment. The parallel efficiency of the distributed approach, however, is critically dependent on the duration of the inherently serial master phase relative to that of the bottleneck subproblem. This paper investigates strategies for improving efficiency in distributed Dantzig—Wolfe decomposition by better balancing the load between the master and subproblem processors. We report computational experience on an Intel iPSC/2 hypercube multiprocessor with test problems having dimensions up to about 30 000 rows, 87 000 columns, and 200 coupling constraints.
Similar content being viewed by others
References
G.M. Amdahl, “The validity of the single-processor approach to achieving large-scale computer capabilities,”AFIPS Conference Proceedings 30 (1967) 483–485.
V. Chvatal,Linear Programming (Freeman, New York, 1983).
G.B. Dantzig and P. Wolfe, “The decomposition principle for linear programs,”Operations Research 8 (1960) 101–111.
D. De Witt, R. Finkel and M. Solomon, “The CRYSTAL multicomputer: design and implementation experience,”IEEE Transactions on Software Engineering SE-13 (1987) 953–966.
Y. Dirickx and L.P. Jennergren,Systems Analysis by Multilevel Methods (Wiley, New York, 1987).
D. Finon, “Un modèle énergétique pour la France,” Editions du CNRS, Paris (1976).
K.C. Gilbert, W.A. McKenney and R.S. Garfinkel, “An algorithm for capacitated material requirements planning,” Paper presented at theORSA/TIMS Joint National Conference, Denver, CO (October 1988).
J.L. Gustafson, “Reevaluating Amdahl's Law,”Communications of the ACM 31 (1988) 532–533.
J.P. Hayes, T.N. Mudge, Q.F. Stout, S. Colley and J. Palmer, “Architecture of a hypercube supercomputer,”Proceedings of the International Conference on Parallel Processing (1986) 653–660.
J.K. Ho, “Decomposition of linear programs using concurrent processing on multicomputers,” Final Report, RP1999-11, Electric Power Research Institute (1989).
J.K. Ho, T.C. Lee and R.P. Sundarraj, “Decomposition of linear programs using parallel computation,”Mathematical Programming 42 (1988) 391–405.
J.K. Ho and E. Loute, “An advanced implementation of the Dantzig—Wolfe decomposition algorithm for linear programming,”Mathematical Programming 20 (1981) 303–326.
J.K. Ho and E. Loute, “Computational experience with advanced implementation of decomposition algorithms for linear programming,”Mathematical Programming 27 (1983) 283–290.
J.K. Ho and E. Loute, “Computational aspects of DYNAMICO: a model of trade and development in the world economy,”Revue Française d'Automatique, Informatique et Recherche Opérationelle 18 (1984) 403–414.
J.K. Ho and R.P. Sundarraj,DECOMP: an Implementation of Dantzig—Wolfe Decomposition for Linear Programming, Lecture Notes in Economics and Mathematical Systems No. 338 (Springer, New York, 1989).
IBM Corporation, “MPSX/370 Version 2 program reference manual,” SH19-6553 (1988).
Intel Corporation, Beaverton, OR, “iPSC/2 Green Hills Fortran language reference manual (preliminary),” No. 311020-003 (1988).
Intel Corporation, Beaverton, OR, “iPSC/2 user's guide,” No. 311532-004 (1989).
C. Lemarechal, J.J. Strodiot and A. Bihain, “On a bundle algorithm for nonsmooth optimization,” in: OL. Mangasarian, R.R. Meyer and S.M. Robinson, eds.,Nonlinear Programming 4 (Academic Press, New York, 1981) pp. 245–282.
D. Medhi, “Parallel bundle-based decomposition for large-scale structured mathematical programming problems,”Annals of Operations Research 22 (1990) 101–127.
J.L. Nazareth,Computer Solution of Linear Programs (Oxford University Press, New York, 1987).
J.B. Rosen and R.S. Maier, “Parallel solution of large-scale, block-angular linear programs,”Annals of Operations Research 22 (1990) 23–41.
R.P. Sundarraj and J.K. Ho, “Distributed simplex algorithms for linear programming”, Working Paper No. 92-103, Graduate School of Management, Clark University (Worcester, MA, 1992).
Author information
Authors and Affiliations
Additional information
This paper is dedicated to Phil Wolfe on the occasion of his 65th birthday.
Rights and permissions
About this article
Cite this article
Gnanendran, S.K., Ho, J.K. Load balancing in the parallel optimization of block-angular linear programs. Mathematical Programming 62, 41–67 (1993). https://doi.org/10.1007/BF01585159
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF01585159