The part of the genetic algorithm that takes up most of the time is
the evaluation of the fitness function. The number of evaluations of
the fitness in each generation depends on the operators used and the
population update model. Different operators and update models can
lead to very different numbers of evaluations per generation. That is
the reason why it is common to use the number of evaluations as the
stop criterion instead of the number of generations. We have used a
limit of 300,000 evaluations [Evv98,dK98] as stop
criterion. The precision of the solutions is bounded by the precision
of the data type used in the implementation of the genetic algorithm.
We have used a double precision data type of 64 bits following the
specification ANSI/IEEE STD 754-1985 (IEEE Standard for Binary
Floating-Point Arithmetic). This data type has a precision of 15 - 17
digits.

Domingo
2005-07-11