On solving large maximum concurrent flow problems

  • Authors:
  • Farhad Shahrokhi;David W. Matula

  • Affiliations:
  • Southern Methodist University;Southern Methodist University

  • Venue:
  • CSC '87 Proceedings of the 15th annual conference on Computer Science
  • Year:
  • 1987

Quantified Score

Hi-index 0.00

Visualization

Abstract

The maximum concurrent flow problem (MCFP) is readily illustrated by problems such as traffic flow in road networks and message transfer in packet switched networks. The MCFP was introduced in [MA85] and can be solved using either linear programming techniques or using the flow routing algorithms [BM86, TM86]. Application of these techniques has typically been limited to problems of small size (40 vertices or less). We present a refinement of the algorithm in [BM86] which operates successfully on problems three times larger than those solved by the previous techniques. Empirical polynomially bounded storage and time complexity is demonstrated and applications in packet switched networks are investigated.