Friday, March 14

4:00 PM-6:00 PM
Nicollet D3

Parallel Linear Programming: Practical Approaches and Implementations

Linear programming problems are pervasive in science, engineering, and industry. Modern mathematical models often lead to optimization problems with many thousands of variables and with stochastic or discrete characteristics. Moreover, these modeling and optimization activities often are part of a design process that requires rapid turnaround and repeated trials. In this context, effective and robust parallel methods are essential.

Despite the enormous advances in large-scale linear programming algorithms, relatively little attention has been given to the development of practical parallel approaches.

This minisymposium will present recent work in parallel algorithms and software for large-scale linear programming: continuous, stochastic, and discrete problems. Emphasis will be on performance and implementation issues, including the important role played by parallel sparse linear algebra.

Organizers: Thomas F. Coleman, Cornell University and Stephen J. Wright, Argonne National Laboratory

4:00 Fully Parallel Generic Branch-and-Cut
Márta Esö, Cornell University; László Ládanyi and Theodore K. Ralphs, U.S. Air Force, Nellis Air Force Base; and Leslie E. Trotter, Jr. , Cornell University
4:30 Graph-Partitioning Based Sparse Matrix Orderings for Interior-Point Algorithms
Anshul Gupta, IBM T. J. Watson Research Center
5:00 Scalable Implementation of Nested Decomposition for Stochastic Programming
Steven E. Wright, Miami University, Oxford, Utah
5:30 pPCx: A Parallel Interior Predictor-Corrector Method for Linear Programming: Implementation and Performance
Thomas F. Coleman, Organizer; Joseph Czyzyk, Chunguang Sun, Michael Wagner, Cornell University, and Stephen J. Wright, Organizer

PP97 Homepage | Program Updates|
Registration | Hotel Information | Transportation | Program Overview | Program-at-a-Glance

MMD, 3/5/97