GEP Book

  Home
  News
  Author
  Q&A
  Tutorials
  Downloads
  GEP Biblio
  Contacts

  Visit Gepsoft

 

C. FERREIRA Invited Tutorial Presented at WSC6, 2001

Gene Expression Programming in Problem Solving

Gene Transposition
 

In gene transposition an entire gene functions as a transposon and transposes itself to the beginning of the chromosome. In contrast to the other forms of transposition, in gene transposition, the transposon (the gene) is deleted at the place of origin.

Apparently, gene transposition is only capable of shuffling genes, and for ETs linked by commutative functions, this contributes nothing to adaptation in the short run. However, gene transposition is very important when coupled with other operators (all kinds of GEP recombination; see next section), for it allows not only the duplication of genes but also a more generalized recombination of genes or smaller building blocks.

The chromosome to undergo gene transposition is randomly chosen, and one of its genes (except the first, obviously) is randomly chosen to transpose. Consider the following chromosome composed of 3 genes:

012345678901201234567890120123456789012
/+Qa*bbaaabaa*a*/Qbbbbbabb/Q-aabbaaabbb

Suppose gene 3 was chosen to undergo gene transposition. Then the following chromosome is obtained:

012345678901201234567890120123456789012
/Q-aabbaaabbb/+Qa*bbaaabaa*a*/Qbbbbbabb

Note that for numerical applications where the function chosen to link the genes is commutative, the expression evaluated by the chromosome is not modified. But the situation differs in other applications where the linking function is not commutative, for instance, the IF function chosen to link some sub-ETs in Boolean problems (see Figure 3). Note that, in this case, gene transposition has a very drastic effect, generating most of the times nonviable individuals.

Home | Contents | Previous | Next