Experiments with parallel branch-and-bound algorithms for the set covering problem

Authors:
Russell A. Rushmeier;George L. Nemhauser
Affiliations:
USAir DCA/H700, 2345 Crystal Drive, Arlington, VA 22227, USA;School of Industrial and Systems Engineering, Georgia Institute of Technology, Atlanta, GA 30332, USA
Venue:
Operations Research Letters
Year:
1993

Citing 6
Cited 1

An introduction to parallelism in combinatorial optimization

Discrete Applied Mathematics
Integer and combinatorial optimization

Integer and combinatorial optimization
Optimal solution of set covering/partitioning problems using dual heuristics

Management Science
Large-scale 0-1 linear programming on distributed workstations

Annals of Operations Research
Parallel algorithms for the set covering problem

Parallel algorithms for the set covering problem
Results from a parallel branch and bound algorithm for the asymmetric traveling salesman problem

Operations Research Letters

Simulating realistic set covering problems with known optimal solutions

Computers and Industrial Engineering

Quantified Score

Hi-index	0.00

Visualization

Abstract

We discuss the results of computational experiments using a new parallel multi-task model for integer optimization. The model is implemented in C on a BBN TC2000 computer. Eight branch-and-bound algorithms based on the use of two types of tasks, linear programming and subgradient optimization with Lagrangian relaxation, are tested on difficult randomly generated set covering problems. We investigate the influence of relaxation choice and node selection strategy on parallel performance. Results indicate that the Lagrangian relaxation using a mixed selection strategy is effective on the largest problems. The best overall algorithms divide the computer resources into two distinct searches, which communicate only to update global information on lower and upper bounds.