Paper, SIGAda, November 14–18, 2004, Atlanta, Georgia, USA.
Genetic Algorithms provide computational procedures that are modeled on natural genetic system mechanics, whereby a coded solution is evolved from a set of potential solutions, known as a population. GAs accomplish this evolutionary process through the use of basic operators, crossover
and mutation. Both the representation of the population and the operators require careful scrutiny, and can change dramatically for different classes of problems. Initial tests were conducted using a GA written in Ada95, and required substantial modifications to handle the changing domains.
Subsequent testing was done with a toolbox constructed for Matlab, but the class of problems it can solve is restrictive. Ada95’s generic mechanism for parameterization would allow
for reuse of existing structures for a broader range of problems. This paper describes the tests performed thus far using both approaches, and compares the performance of the two approaches with regards to optimization.
