基于项目网络支配和扩散关系的研发项目组合选择

项目组合包含多项目且项目间存在相互作用和依赖关系，针对传统项目组合评价方法忽略了各项目间依赖关系的不足，本文采用复杂网络理论和PageRank算法，提出一种新的项目优先级排序方法（PPRM）。首先，本文建立研发项目多属性评价准则，分析了项目间的支配关系以及技术和经验在项目间的扩散关系。然后，以项目为节点、以支配和扩散关系为边，分别构建了项目支配和扩散网络，进一步，采用设计结构矩阵（DSM）和K-shell方法建立了基于支配网络的项目影响力模型，并建立了考虑项目之间多次扩散传播的综合扩散概率模型。综合项目节点影响力和扩散关系，本文构建了基于PageRank算法的研发项目优先级排序模型。最后，以某研发项目组合选择为例，验证了本文所建立的模型和算法可有效分析项目组合中的排序问题。

R&D Project Portfolio Selection Based on Domination and Diffusion Relationship in the Project Network

To gain competitive leverage, firms must seek to continuously invest in a large number of research and development (R&D) projects. The R&D project portfolio contains many projects and various interdependencies. However, existing methods of portfolio selection usually don't take these interdependencies into account. Using the complex network theory, the PageRank and K-shell algorithm, a new project priority ranking method (PPRM) is propozed in this paper. The following key problems in the R&D project portfolio selection need to be solved:1) how to determine multiple criterions for prioritizing projects in the portfolio; 2) how to analyze the interaction relationship among projects via the complex network theory; 3) how to measure projects' priority by considering both domination and diffusion relationship via the K-shell and PageRank algorithm.
Hence, the interaction and interdependency among projects in the R&D project portfolio selection from the perspective of domination and technical diffusion relationship are focused on. First, a multi-attribute evaluation criterion is established to analyze the domination and diffusion relationship among projects. Then, the domination and diffusion networks of the R&D portfolio are built. Further, using the design structure matrix (DSM) and K-shell method, the network node's influencing model of the project is built via the domination network and the probability model of multiple diffusion propagation is also built via the diffusion network. Considering the influence and diffusion relationship of each project, the model of project's priority ranking is present based on the improved PageRank algorithm. Finally, an industrial example is provided to illustrate the proposed models. The results yield and reinforce several managerial insights, including:1) both the relationship of domination and technical diffusion among projects have a obvious influence on the project priority ranking; 2) the priority of each project may be improved by enhancing its domination or diffusion ability, therefore, the project manager should identify which reasons (i.e., diffusion or dominance relationship) will lead to the change of the project priority and understand the change trend, then realize a trade-off analysis of project portfolio selections.