Common feature techniques for discrete optimization

  • Authors:
  • A. J. Goldstein;A. B. Lesk

  • Affiliations:
  • -;-

  • Venue:
  • DAC '76 Proceedings of the 13th Design Automation Conference
  • Year:
  • 1976

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present a general approach to finding optimal arrangements of objects, given a cost function for evaluating an arrangement. The method is based on the assumption that features which are common to many weak local optima of a problem should be present in the global optimum. The algorithm identifies such common features and uses them to create “blocks” of objects which are treated as indivisible units. We have used general-purpose algorithms which do not exploit the peculiarities of any one problem. Thus, the method described here may not be as good as an heuristic which has been tailored to a particular problem. However, it is easily adaptable to different problems, and produces many near-optimal solutions. Two examples are discussed: an electrical-net wiring problem and the traveling salesman problem.