Abstract: The evaluation of complex surface profile error is usually difficult to perform in the field of ultra-precise manufacture and measurement. On the basis of analysis of study status of complex surface profile error, the key problem is proposed, the definition of complex surface profile error is description, and its math model is established. On the basis of analysis of NURBS surface, the subdivision approach algorithm for calculating the minimum distance between point and surface is proposed. The contradiction among calculation accuracy, encoding length and computation complexity of standard genetic algorithm is analyzed, the improved genetic algorithm is proposed, its crossover operator and mutation operator are offered. The detailed steps are established for calculating complex surface profile error based on subdivision approach algorithm and genetic algorithm. It can obtain precision result to calculate complex surface profile error by use of genetic algorithm with canonicity real number encoding and subdivision approach algorithm. This method can be realized easily on computer and is very suitable for three coordinate measuring machine.

Key words: Data processing, Complex surface profile error, Genetic algorithm