Abstract: If we only focus on the project attributes without considering the impact of the project relevance and the project-derived technology, experience/information diffusion on the project portfolio decisions, it would lead to decision bias, that is, it would underestimate the value of potentially technology-leading projects. To avoid this issue, this paper uses the complex network theory to consider the domination and diffusion relationship between projects as directed weighted network. The K-shell decomposition approach is used to build models in a portfolio network(the project impact model based on dominance relationship or the model of technology, experience/information diffusion among projects, respectively). Then, based on the PageRank, the project priority ranking decision model is established by considering the relationship between project dominance and diffusion. To evaluate the feasibility of the proposed model and algorithm, a numerical study is conducted and provides a reference for enterprises’ project portfolio decision.

Key words: Portfolio decision, Design structure matrix(DSM), PageRank algorithm, K-shell algorithm

摘要： 单纯侧重项目自身属性而不考虑项目关联性以及由项目衍生而来的技术、经验/信息扩散对项目组合决策时的影响,易导致决策偏差,低估具有潜在技术先导性项目的价值。对此,引用复杂网络理论,以项目关联性的视角,将项目间支配和扩散关系分别抽象为有向加权网络,运用K-shell分解方法构建项目组合网络中基于支配关系的项目影响力模型以及技术、经验/信息在项目间扩散传播的模型。然后,基于PageRank算法,综合考虑项目间支配与扩散关系,建立了项目优先级排序决策模型。最后,通过算例分析说明了该模型与算法的可行性与有效性,为企业项目组合决策提供了有益的参考。

关键词: 项目组合决策, 设计结构矩阵(DSM), PageRank算法, K-shell算法