Domain Information Specification

The DARWIN language provides support for domain specification of the simple types data members of a genetic construct. By specifying a domain, the compiler is informed about the legal values for the given data member and it will generate necessary code in the genetics operators which will guarantee that no domain violations are resulting. This feature is extremely useful because the compiler generated initializer, crossover and mutation function will not be generating illegal solutions, thus resulting in more intelligent and robust implementations.

The syntax for specifying domains is as follows:

domain	::=	"$<$" "[" elements "]"
	$\vert$	"$<$" "{" elements "}"
	$\vert$	/* no domain */

Currently, the DARWIN language supports just the range and set domains. The range domain specifies the upper and lower boundaries of acceptable values and the set domain lists the possible values for a construct member.

For example, is a construct member x is of type char and can only take values in the range [0..100], then it can be declared as follows:

  char x < [0..100];