GEP Book

  Home
  News
  Author
  Q&A
  Tutorials
  Downloads
  GEP Biblio
  Contacts

  Visit Gepsoft

 

C. FERREIRA Complex Systems, 13 (2): 87-129, 2001

Gene Expression Programming: A New Adaptive Algorithm for Solving Problems

Selection
 

In all the problems presented in this work, individuals were selected according to fitness by roulette wheel sampling [10] coupled with the cloning of the best individual (simple elitism). A preliminary study of different selection schemes (roulette wheel selection with and without elitism, tournament selection with and without elitism, and various kinds of deterministic selection with and without elitism) suggests that there is no appreciable difference between them as long as the cloning of the best individual is guaranteed (results not shown). Some schemes perform better in one problem, others in another. However, for more complex problems it seems that roulette wheel selection with elitism is best.

Home | Contents | Previous | Next