Research Output

Enhancing the performance of a GA through visualisation.

  This article describes a new tool for visualising
genetic algorithms, (GAs) which is designed
in order to allow the implicit mechanisms
of the GA | i.e. crossover and mutation
| to be thoroughly analysed. This
allows the user to determine whether these
mechanisms are essential to a GAs performance,
and if so, to provide a principled
means of setting the parameters associated
with them, based on a sound understanding
of their e�ects. The use of the tool is illustrated
by applying to the analysis of a jobshop
scheduling problem, in order to choose
e�ective operators, and to determine appropriate
settings for them. We show that by
analysing two crossover operators and a mutation
operator, we can re�ne the choice and
settings of these parameters in order to improve
the performance of the GA on the particular
problem chosen. When the new operators
are applied to a wider range of problems
of the same type, a similar improvement
in performance is observed.

  • Type:

    Book Chapter

  • Date:

    30 November 1999

  • Publication Status:


  • Publisher

    Morgan Kaufmann

  • Library of Congress:

    QA75 Electronic computers. Computer science

  • Dewey Decimal Classification:

    006.3 Artificial intelligence


Hart, E. & Ross, P. (1999). Enhancing the performance of a GA through visualisation. In Proceedings of GECCO-2000. Morgan Kaufmann. ISBN 1-55860-708-0



genetic algorithms; visualisation; crossover; mutation; parameters; jobshop scheduling; performance improvement;

Available Documents