GEP Book

  Home
  News
  Author
  Q&A
  Tutorials
  Downloads
  GEP Biblio
  Contacts

  Visit Gepsoft

 

C. FERREIRA In E. Lutton, J. A. Foster, J. Miller, C. Ryan, and A. G. B. Tettamanzi, eds., Proceedings of the 4th European Conference on Genetic Programming, Lecture Notes in Computer Science, Vol. 2278, pages 51-60, Springer-Verlag, Berlin, Germany, 2002.

Discovery of the Boolean Functions to the Best Density-Classification Rules Using Gene Expression Programming

Discovering the Boolean Functions to the Best Rules for the Density-Classification Task
 
The space of possible rules for Boolean functions of seven arguments is the huge space ofrules, and the size of the space of possible computer programs that can be composed using the elements of the function and terminal sets is greater still. Therefore, the discovery of the Boolean functions to the Coevolution1, and Coevolution2 rules is no trivial task and, in fact, their discovery by GEP involved several optimization runs where the best solution of a run was used as seed to evolve better programs. This kind of strategy is inevitable while trying to solve real-world problems of great complexity. Good intermediate solutions are some times hard to find, and start everything anew is no guarantee that a perfect solution or even a better intermediate solution will be found. Therefore, it is advantageous to use a good intermediate solution as seed to start a new evolutionary cycle, with the advantage that the seed or the evolutionary conditions can be slightly changed before an optimization. For instance, we can increase the gene length, introduce or remove a neutral gene, introduce or remove new functions in the function set, linearize a multigenic solution, change the fitness function or the selection environment, change the population size, change the genetic operators or their rates, and so forth. The particular evolutionary strategy followed in each case is given below.

Home | Contents | Previous | Next