Abstract: In view of the complexity of design process planning problem and limitation of existed methods, this paper takes total time and cost of the whole tasks in product development as objectives by considering all kinds of uncertain factors in practical product development and further described design process planning as a simulation based stochastic optimization problem. A new hybrid algorithm is proposed to solve this problem by introducing optimal computing budget allocation technique into genetic algorithm framework to improve algorithm searching efficiency and result reliability. Finally, the development of rotor and bearing system in turbine is adopted as example to validate the effectiveness of proposed method. Results demonstrate effectiveness of modeling method and high efficiency of solving algorithm. The method can be extended to solve all kinds of product development process and is of universality.

Key words: design process planning, design iteration, stochastic optimization, hybrid optimization algorithm, rotor and bearing system

摘要： 针对目前产品设计过程规划研究中存在的不足，在充分考虑实际设计过程中存在的各种不确定因素的基础上，以产品开发过程中的全体任务为规划对象，以设计迭代时间和成本为目标，将设计过程规划问题描述为基于仿真的随机优化问题进行处理。提出一种模型求解的混合遗传算法，该算法引入最优计算量分配技术进行样本分配，极大地提高了算法的搜索效率，有效地改善了遗传算法搜索的可靠性。以汽轮机轴承转子系统的设计为例，对提出的方法的有效性进行了验证，仿真结果表明，该建模方法有效且算法求解效率高。该方法可推广应用于各种产品设计过程的规划，具有普遍意义。

关键词: 过程规划, 设计迭代, 随机优化, 混合优化算法, 轴承转子系统