Parametric Analysis for Generalized Network Flow Problems.

Abstract

A generalized network flow problem can be efficiently solved with state of the art computer codes. However, because of the uncertainty of inaccuracy of initial problem input data, the solution found may need reexamination. Parametric analysis allows the management scientist to vary any of the right-hand side restrictions, either external flow or capacities, and determine the series of bases so the solution remains optimal. This study uses the special structural properties of the generalized network flow problem to iteratively change the network flows, until no further changes in the basis can be made. Additionally, two dual-incremental flow algorithms based on parametric analysis are developed. A series of test problems were randomly generated and computational results were compared to two primal generalized network computer codes and a published dual incremental computer code. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 01, 1984
Accession Number
ADA145558

Entities

People

  • M. E. Baum

Organizations

  • Air Force Institute of Technology

Tags

Communities of Interest

  • Cyber

DTIC Thesaurus Topics

  • Air Force
  • Algorithms
  • Computational Science
  • Computations
  • Computer Programming
  • Computers
  • Families (Human)
  • Linear Programming
  • Linear Systems
  • Literature Surveys
  • Mathematical Programming
  • Operations Research
  • Optimization
  • Parametric Analysis
  • Parametric Programming
  • Simplex Method
  • Terminals

Fields of Study

  • Computer science

Readers

  • Computational Modeling and Simulation
  • Operations Research
  • Theoretical Analysis.