Given a real-based or permutation-based function, and the
associated search space, gaoptim will perform a
function maximization using the Genetic Algorithm
approach. For better performance, a real-number encoding
is used.
All you need to get started is to provide a function and
the associated search space - there are sensible defaults
to all the other parameters. On the other hand, you can
provide custom genetic-operators to control how your
population will reproduce and mutate (see
the examples).
After setting the algorithm parameters, you can evolve
your population and check the results. You don't need to
do this in one step, you can always evolve a small number
of generations and query the best solution found. If this
solution doesn't fit your needs, you can keep evolving
your population - this approach saves time and computer
resources.
References
Randy L. Haupt, Sue Ellen Haupt (2004). Practical genetic
algorithms - 2nd ed.