A Connectionist Machine for Genetic Hillclimbing

  • David H. Ackley

Table of contents

  1. Front Matter
    Pages i-xiii
  2. David H. Ackley
    Pages 1-28
  3. David H. Ackley
    Pages 29-70
  4. David H. Ackley
    Pages 71-102
  5. David H. Ackley
    Pages 103-131
  6. David H. Ackley
    Pages 133-153
  7. David H. Ackley
    Pages 155-189
  8. David H. Ackley
    Pages 191-201
  9. David H. Ackley
    Pages 203-230
  10. Back Matter
    Pages 231-260

About this book


In the "black box function optimization" problem, a search strategy is required to find an extremal point of a function without knowing the structure of the function or the range of possible function values. Solving such problems efficiently requires two abilities. On the one hand, a strategy must be capable of learning while searching: It must gather global information about the space and concentrate the search in the most promising regions. On the other hand, a strategy must be capable of sustained exploration: If a search of the most promising region does not uncover a satisfactory point, the strategy must redirect its efforts into other regions of the space. This dissertation describes a connectionist learning machine that produces a search strategy called stochastic iterated genetic hillclimb­ ing (SIGH). Viewed over a short period of time, SIGH displays a coarse-to-fine searching strategy, like simulated annealing and genetic algorithms. However, in SIGH the convergence process is reversible. The connectionist implementation makes it possible to diverge the search after it has converged, and to recover coarse-grained informa­ tion about the space that was suppressed during convergence. The successful optimization of a complex function by SIGH usually in­ volves a series of such converge/diverge cycles.


algorithms control control algorithm genetic algorithms knowledge representation learning memory modeling optimization search strategy supervised learning

Authors and affiliations

  • David H. Ackley
    • 1
  1. 1.Carnegie Mellon UniversityUSA

Bibliographic information

  • DOI
  • Copyright Information Springer-Verlag US 1987
  • Publisher Name Springer, Boston, MA
  • eBook Packages Springer Book Archive
  • Print ISBN 978-1-4612-9192-3
  • Online ISBN 978-1-4613-1997-9
  • Series Print ISSN 0893-3405
  • Buy this book on publisher's site