LAST UPDATED: July 20, 1996

Fourteen "standard" sample applications are provided --

  • Goldberg's three example problems, called here app, app1, and app2, as described in his book.
  • Holland's Royal Road "Challenge" problem, as posed at ICGA '93. Also demos fairly exotic shell script (or bat file) for running GALOPPS in a dynamic, hierarchical fashion, while minimizing any unnecessary function evaluations for population initialization.
  • Four of DeJong's five test functions.
  • A blind traveling salesperson problem.
  • A demo of the non-binary alphabet representation and operators.
  • A demo of the subpopulation-level inversion capability.
  • A demo of the capability to run subpopulations using different representations, with migration among them (supporting the "injection island" architecture and similar capabilities).
  • A demo of the use of fields of different lengths, with GALOPPS automatically tracking the fields (works even under inversion operations -- see also the inversion demonstration example, which also uses different-length fields).
  • A demo of Emanuel Falkenauer's "Grouping Genetic Algorithm (GGA)", on a bin packing example (classical 1-D bin packing). Code is generic GGA code, written for GALOPPS by Brian Zulawinski, so the operators can be used for other applications of the GGA (only the file appbpp.c is specific to this problem), but the code is not yet accompanied by a "user's guide", so, although commented, may be difficult to follow.

(Note: the "hybrid" rep (mixture of permutation and values encoded) problems have been dropped from the standard distribution in Release 3.0. Information is available to interested parties from the author.)

Galopps has been used for many other applications, both in the MSU GARAGe and around the world, but they are specialized &/or proprietary, and are not distributed with galopps3.2.

For more information on GALOPPS, contact Erik Goodman

GALOPPS, Erik Goodman