Using simulated annealing for knot placement for cubic spline approximation

In this paper, a new methodology is developed for knots placement for cubic spline approximation, using Simulated Annealing. It is not necessary to convert the problem into a discrete combinatorial optimization problem, as presented in other paradigms inspired in genetic algorithm or artificial immune systems, and therefore, removing the constrain of combinatorial optimization problem in the proposed methodology, because the real value of the location of the knots is directly optimized. The accuracy, computationally efficient and robustness of the algorithm presented is compared by different experimental result, with other approaches presented in the bibliography.