Knot-removal surface fairing using search strategies

This paper presents two automatic fairing algorithms for parametric C-continuous bi-cubic B-spline surfaces. The fairing method consists of a knot removal and knot reinsertion step which locally smooths the surface. Search strategies like best-first-search and simulated-annealing are searching for the global minimum of the fairing measure. The best-first-search algorithm constructs only partially a search tree and reduces significantly the complexity of a systematic search. Simulated annealing is a heuristic algorithm which needs a probability function and some further parameters as input. Both methods can satisfy end constraints and tolerances. Their performance is discussed for two numerical experiments.