Abstract: To get the automation of model partitioning, reduce the influence of human factors to partitioning results, increase the precision and efficiency of computation and ensure the rationality of partitioning scheme, a feature-based partitioning algorithm for complex model is proposed. Model partitioning schemes are gained through four main steps：feature recognition, manufacturability analysis, partitioning faces abstraction and multi-objective optimization of partitioning schemes. For there are many objective functions in partitioning process, multi-objective optimization of genetic algorithm is applied in algorithm. The optimal partitioning scheme is obtained by partitioning face and partitioning sequence optimization and the feasibility of this algorithm is verified by concrete examples.

Key words: Feature-based, Genetic algorithm, Manufacturability analysis, Model partitioning manufacturing, Multi-objective optimization