Automatic knot adjustment by an improved genetic algorithm

In order to obtain a good B-spline contour model from scattered data, the knots can be respected as variables. A curve is then modeled as a continuous, nonlinear and multivariate optimization problem with many local optima. To overcome the defects of traditional genetic algorithm in adjusting knots, an improved genetic algorithm is designed in this paper. In order to improve its searching space and convergence performance, the improved genetic algorithm adopts float-coding and introduces a dynamic adaptive strategy to adjust the crossover rate (P_c) and mutation rate (P_m). Results show that the improved genetic algorithm maintains the population diversity and alleviates the problem of premature convergence more effectively, and determines a more appropriate location of knots.