COMPUTATIONAL EXPERIENCE WITH A MULTICOMMODITY NETWORK FLOW ALGORITHM

This paper discusses the computational results achieved with an experimental implementation of an algorithm for solving the minimum cost multicommodity flow problem for directed graphs. As programmed, the method combines aspects of network flow, partition programming, working bases and the use of secondary constraints. Computationally, this approach is particularly attractive within the multicommodity framework. The experimental results reveal the efficiency of the algorithm as solution strategies and problem parameters are varied. A comparison is also made with the corresponding linear programming solutions.

Media Info

  • Features: References;
  • Pagination: p. 205-226

Subject/Index Terms

Filing Info

  • Accession Number: 00096604
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Jul 29 1975 12:00AM