MetaCart Sign in to MyCiteSeerX

Include Citations | Advanced Search | Help

Disambiguated Search | Include Citations | Advanced Search | Help

A Taxonomy of Control Schemes for Genetic Code Growth (1997) [1 citations — 0 self]

by Byoung-tak Zhang
Michigan State University, East Lansing, Michigan
Add To MetaCart

Abstract:

One important issue in evolutionary computation with variable size representation is the control of code growth. This paper identifies four distinctive forms of parsimony pressure useful for growth control. A taxonomy of control schemes is presented which provides a guideline for the comparison of existing methods and the development of novel methods. Based on a taxonomical analysis of widespread methods we suggest some promising areas of further research. 1 Introduction In evolutionary computation with variable size representations, such as genetic programming, code complexity tends to grow as generation goes on. There are, however, several reasons why simpler and parsimonious codes should be preferred. First, large codes are more difficult to analyze and understand. Second, complex codes require more computer resources in space and time for their evolution and execution. Third, the generality of codes tends to decrease as their size increases. The last point is related with the prin...

Citations

1921 Genetic Programming I : On the Programming of Computers by Means of Natural Selection – Koza - 1992
77 Coevolving high-level representations – Angeline, Pollack - 1994
60 Balancing accuracy and parsimony in genetic programming – Zhang, Mühlenbein - 1995
46 Discovery of subroutines in genetic programming – Rosca, Ballard - 1996
26 Evolving optimal neural networks using genetic algorithms with Occam’s razor – Zhang, Mühlenbein - 1993
18 System identification using structured genetic algorithms – Iba, Kurita, et al. - 1993
14 Simultaneous discovery of reusable detectors and subroutines using genetic programming – Koza - 1993
1 Generality and difficulty in genetic programming – Kinnear - 1993