ThesisGeometric modeling in Computer-Aided-Design by multi-objective optimization is very common nowadays. The solutions focus in minimizing the approximation error from a given set of points, fitting and fairing. To solve difficult multimodal optimization problems, Genetic Algorithm have been developed in computer program for control the shape of non-uniform B-spline automatically.
Genetic Algorithm construct individuals by considering candidates of the locations of knots as genes, search for the best model among the candidate models by using Bayes Information Criterion (BIC). The evolution of modeling has been developed with this aim.