Abstract: According to the dynamic information of manufacture resource, the process plan system should make process plan decision accurately and timely for e-manufacture. The process plan includes three stages, part information input, feature process selection and process plan decision-making. Error back propagation neutral network algorithm has been used to solve feature process chains fuzzy match. The dynamic information of resources and rules of process plan are regarded as the integrated restriction conditions, and the genetic algorithm is adopted to realize the process route optimization selection based on the dynamic resource information. It is effective to solve resource conflict in production schedule. Finally, the feasibility of this design method is proved by a case running on the prototype that is developed by Java.

Key words: E-manufacture, Dynamic process plan, Back propagation algorithm, Genetic algorithm

摘要： e-制造工艺系统要求及时准确地根据资源动态信息进行工艺路线的实时决策。该文将工艺规划过程划分为零件信息输入、特征工艺选择和工艺路线的决策等3个阶段，采用神经网络BP算法解决了特征工艺链模糊匹配问题，将资源动态信息与工艺规则共同作为约束条件，采用GA算法实现基于资源动态信息的工艺路线优化选择，有效地解决了调度中的资源冲突问题。采用Java技术开发了软件原型系统，通过实例分析验证了设计方法的可行性。

关键词: e-制造, 动态工艺规划, BP算法, 遗传算法