We have already met two different domains in GEP genes: the head and the tail. And now another one – the Dc domain – will be introduced. This domain was especially designed to handle random numerical constants and consists of an extremely elegant, efficient, and original way of dealing with them.

As an example, Genetic Programming handles random numerical constants by using a special terminal named “ephemeral random constant” (Koza 1992). For each ephemeral random constant used in the trees of the initial population, a random number of a special data type in a specified range is generated. Then these random constants are moved around from tree to tree by the crossover operator. Note, however, that with this method no new constants are created during evolution.

Gene Expression Programming handles random numerical constants differently (Ferreira 2001). GEP uses an extra terminal “?” and an extra domain Dc composed of the symbols chosen to represent the random constants. The values of each random constant, though, are only assigned during gene expression. Thus, for each gene, the initial random constants are generated during the creation of the initial population and kept in an array. However, a special operator is used to introduce genetic variation in the available pool of random constants by mutating the random constants directly. In addition, the usual operators of GEP (mutation, inversion, transposition, and recombination), plus a Dc-specific transposition and a Dc-specific inversion, guarantee the effective circulation of the random constants in the population. Indeed, with this scheme of constants’ manipulation, the appropriate diversity of numerical constants is generated at the beginning of a run and maintained easily afterwards by the genetic operators. Let’s take then a closer look at the structural organization of this extra domain and how it interacts with the sub-ET encoded in the head/tail structure.

Structurally, the Dc comes after the tail, has a length equal to t, and is composed of the symbols used to represent the random constants. Therefore, another region with defined boundaries and its own alphabet is created in the gene.

Consider the single-gene chromosome with an h = 8 (the Dc is shown in blue):

where the terminal “?” represents the random constants. The expression of this kind of chromosome is done exactly as before, obtaining:

Then the ?’s in the expression tree are replaced from left to right and from top to bottom by the symbols (numerals, for simplicity) in Dc, obtaining:

The random constants corresponding to these symbols are kept in an array and, for simplicity, the number represented by the numeral indicates the order in the array. For instance, for the following array of 10 elements:
 

the chromosome (10) above gives:

This type of domain was used to great advantage not only in symbolic regression but also in parameter optimization and polynomial induction (Ferreira 2002a). But this elegant structure can also be used to evolve the weights and thresholds of evolvable neural networks (Ferreira 2004b) and to encode decision trees with numerical attributes (unpublished material). And we will see here for the first time that this kind of domain can also be used to create Automatically Defined Functions with random numerical constants.