GEP Book

  Home
  News
  Author
  Q&A
  Tutorials
  Downloads
  GEP Biblio
  Contacts

  Visit Gepsoft

 

C. FERREIRA In H. J. Caulfield, S.-H. Chen, H.-D. Cheng, R. Duro, V. Honavar, E. E. Kerre, M. Lu, M. G. Romay, T. K. Shih, D. Ventura, P. P. Wang, Y. Yang, eds., Proceedings of the 6th Joint Conference on Information Sciences, 4th International Workshop on Frontiers in Evolutionary Algorithms, pages 614-617, Research Triangle Park, North Carolina, USA, 2002.

Mutation, Transposition, and Recombination: An Analysis of the Evolutionary Dynamics

General Settings
 
To make this preliminary analysis of GEP operators, the following relatively simple test function was chosen:

y = a4 + a3 + a2 +a

as it can be exactly solved using relatively small populations and relatively short evolutionary times and also because it is appropriate to study all the genetic operators, including operators specific of multigenic systems like gene recombination.

In all the experiments, a set of 10 random fitness cases chosen from the interval [-10, 10] was used (a values: -4.4229, 9.7485, -1.7641, -7.0436, -6.8656, -8.1246, 5.9982, -0.1057, 5.1629, -0.7231); the fitness function used is based on the absolute error and has a selection range of 100 and a precision of 0.01, giving a maximum fitness of 1000 [3]; the selection was made by roulette-wheel sampling coupled with simple elitism; a population size P of 50 individuals and an evolutionary time G of 50 generations were used; the success rate Ps of each experiment was evaluated over 100 independent runs; F = {+, -, *, /} and T = {a}; and three-genic chromosomes of length 39 linked by addition were used. In the experiments where transposition was switched on, three transposons with lengths 1, 2, and 3 were used. In the evolutionary dynamics, only successful runs were chosen and G was extended to 100 generations so that the dynamics could be better perceived.

Home | Contents | Previous | Next