The genetic algorithm (GA) initial population or constraints
hardness can lead the biased search to be tracked over unfeasible
regions of the solution space. An alternative to overcome this
undesirable situation is the appliance of adaptive or
deterministic technics to adjust some parameters of a GA-optimizer
dedicated to eigenstructure assignment via LQR designs. In this
work, we propose a method for crossover (X-OVER) operation
parameters control based on the population's average fitness and
restrictions satisfability as a reference to adjust those
parameters, in the sense of guiding the search intelligently into
feasible solutions. The proposed method is translated into an
algorithm and is accomplished into a multiobjective genetic
optimizer decision-making unit. Finally, the proposed adaptive
strategy performance is verified into a dynamic systems model.
This research results are presented in two papers, this paper
concerns with the problem formulation and a second paper concerns
with computational simulations and result analysis.
