Combining Exact and Heuristic Approaches for Discrete Optimization

Abstract

In the last decade the computational power of discrete optimization methodology has increased remarkably to the point where problems that could not be solved with days of computation can now be solved in minutes by commercial solvers. This success has stimulated the need for methodology to solve even much larger problems and the desire to solve problems in real-time. We have conducted research that has yielded computationally effective algorithms to provide high-quality solutions to very large-scale planning problems and high-quality solutions in (nearly) real-time to operational problems. Traditionally, this goal has been pursued with heuristic approaches.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Feb 18, 2009
Accession Number
ADA495432

Entities

People

  • George L. Nemhauser
  • Martin W. Savelsbergh

Organizations

  • Georgia Tech Research Corporation

Tags

Communities of Interest

  • Energy and Power Technologies
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Algorithms
  • Coefficients
  • Commodities
  • Computations
  • Computer Programming
  • Evolutionary Algorithms
  • Heuristic Methods
  • Inequalities
  • Integer Programming
  • Linear Programming
  • Operations Research
  • Optimization
  • Test Sets
  • Trees (Data Structures)

Readers

  • Operations Research
  • Systems Analysis and Design