In my problem I plan on solving using GA, chromosomes can be infeasible. Which means, some solutions simply aren't valid and the fitness function cannot be computed for them. My goal is to keep the feasibility status binary (feasible or infeasible) since a quantification of the infeasibility of a chromosome would be terribly complicated or even impossible.
How can this be accomplished and implemented into a genetic algortihm in order to find the best, feasible solution?