부산대학교 도서관

A classical process for quality assessment in man-ufacturing is to perform shape reconstruction of manufactured workpieces for generation of a digital counterpart of the already constructed physical good and further evaluation of the quality of shape. Typically, this process is carried out from a cloud of data points obtained from the workpiece by reverse engineering methods by applying surface reconstructions methods to the point cloud. In this work we assume that the normality of data is not satisfied, and that the point cloud follows other bivariate distribution. In particular, we consider that the approximation function is a combination of univariate exponential and polynomi-als functions, thus leading to a non-convex nonlinear constrained continuous minimization problem. To solve it, our method applies a simplified version of a popular metaheuristic technique called bat algorithm. The method is applied to three illustrative examples of point clouds. The graphical and numerical results show that the method performs quite well and can recover the underlying shape of data with good accuracy.