研发项目是由流程架构、组织架构和产品架构组成的复杂系统.本文首先分析了研发项目复杂性的特征、分类和产生原因,指出设计结构矩阵(DSM)方法是量化分析复杂研发项目的有效工具.进一步,本文从流程架构中迭代和重叠的建模,到流程DSM的排序优化与仿真;从组织架构中团队之间依赖度的测度,到组织DSM聚类准则的确定;从产品架构中部件之间依赖度的测度,到产品DSM的模块化设计以及聚类算法的演变和发展;从单一领域的DSM扩展到多领域矩阵(MDM)的建模与应用,清晰地展示了基于DSM的复杂研发项目建模与优化的几条研究脉络和最新研究进展.最后,本文提出“系统架构的复杂性研究-DSM中元素间依赖关系的测度-DSM优化/聚类”相集成的理论分析框架,并指出了有待进一步研究的关键问题.
R&D projects are complex systems composed by process architecture, organization architecture and product architecture. Firstly, we analyze the features, categories and causes of complexity in R&D projects, and argue that the design structure matrix (DSM) is a quantitative tool to analyze complexity effectively. Further, this paper summarizes the progress in modeling and optimization for complex R&D projects based on DSM and elaborates it in the following frame: from modeling iteration and overlapping in process architecture to sequencing optimization and simulation of process DSM; from measuring de- pendencies between teams in organization architecture to identifying clustering criterion of organization DSM; from estimating dependencies between components in product architecture to modularity design in product DSM, as well as the evolution and development of DSM clustering algorithm; from modeling and application in single domain DSM extending to multi-domain matrix (MDM). Finally, this paper proposes an integrated analysis frame that is exploring complexity of system architecture, measuring dependencies between elements in DSM and optimizing/clustering of DSM. Directions of future research in this field are presented.