复杂产品设计优化的数据与知识管理方法

摘要

为了对复杂产品设计中的设计知识进行有效管理和应用,在前期异构数据管理研究的基础上,提出了跨项目的数据版本管理模型。基于输入输出数据的统计规律建立知识模型,并采用Moore-Penrose广义逆法求解该模型。对已有版本的输入参数进行延拓和抽样以扩大样本点范围,利用获得的知识模型快速求解输出参数,通过综合比较选择优良的输入参数,在设计目标约束下逐步优化输入参数,从而快速收敛到设计目标。实验表明,这种方法能较好地改善设计效率和设计质量。%To realize effective management and application for design knowledge in complex product design process,based on previous studies on heterogeneous design data management,a new cross-project version management model was proposed.A knowledge model was established based on the statistic analysis of the input/output design data,and the Moore-Penrose generalized inverse method was uesd to solve the model.The referenced version data was extended and sampled to extend sample range,and then the output parameters were obtained by the knowledge model.Good input parameters were chose by comprehensive comparison,and input parameters were optimized gradually under the constraints of design requirements,thereby converged at the design objectives rapidly.Experiments showed that this design method improved the design efficiency and design quality.