Visualization of a Simple Genetic
Algorithm for Pedagogical Purposes
Vedrana Vidulin
Bogdan Filipič
Jožef Stefan Institute, Department of Intelligent Systems
[email protected]
[email protected]
• Motivation:
– To facilitate the explanation of how genetic
algorithms work.
• SGA Algorithm:
– Based on the Simple Genetic Algorithm
described in [Goldberg, 1989].
Problem Solved by SGA (1)
364
22 + 2 3 + 2 5 + 2 6 + 28
0 1
0
1
1 0
1 1 0
0
Problem Solved by SGA (2)
• Fitness function
f ( x)  100 * ( x / coeff )n
• Coefficient
coeff  2m  1
• Generation consisted of 10 solutions
• Roulette-wheel selection
Forms of Graphical Representation
Colored Strings
Statistics
Graphical Representation of Best-so-far Fitness
Program Functions
Actions
Inputs
Recommended GA Sources
• A. E. Eiben, J. E. Smith, Introduction to Evolutionary Computing,
Springer, 2003
• Genetic Algorithm – Wikipedia, 2006,
http://en.wikipedia.org/wiki/Genetic_algorithm
• D. E. Goldberg, Genetic Algorithms in Search, Optimization, and
Machine Learning, Addison-Wesley, 1989
• R. L. Haupt, S. E. Haupt, Practical Genetic Algorithms, 2nd Edition,
Wiley-Interscience, 2004
• M. Obitko, P. Slavik, Visualization of genetic algorithms in a learning
environment, Spring Conference on Computer Graphics SCCG'99,
Comenius University, Bratislava, p. 101-106, 1999
• R. E. Smith, D. E. Goldberg, J. A. Earickson, SGA-C: A C-language
Implementation of a Simple Genetic Algorithm, The Clearinghouse for
Genetic Algorithms, Technical Report No. 91002, University of
Alabama, Department of Engineering Mechanics, Tuscaloosa 1994