模糊环境下基于参照依赖的项目组合选择研究

Project portfolio selection, as a strategic management tool, emphasizes how to do the right projects. Most previous studies on project portfolio selection assume decision-makers as completely rationality. Few studies consider behavioral factors of decision-makers, leading to the limited reference value of decision results. We employ fuzzy set theory to describe the project parameters Integrated prospect theory and bounded rationality theory, we proposed a project portfolio selection model based on reference dependence under fuzzy environment. The main contents in our research are as follows: (1) determining the single reference point and the multiple reference point of project portfolio selection problem, and analyze the effect of these reference point to decision-makers' behavior. And, then, the satisfaction degree function is proposed. In addition, we presented the CVaR risk measure for project portfolio under fuzzy environment; (2) Considering ambiguity aversion behavior of decision-makers, a fuzzy multi-objective project portfolio selection model is proposed with the objectives to maximize satisfaction degree, while the ambiguity and the fuzziness of project portfolio benefit are minimized under the realized constraints; (3) according to the features of the model, a hybrid algorithm is designed to solve the presented model; (4) our presented model and algorithms are applied to pharmaceuticals R&D project portfolio selection problems . we compared the proposed model with the traditional model to verify its applicability and effectiveness. The effect of different reference points to the decision results is also studied. This study not only extends the classical project portfolio selection model in theory, but has can provide some guidance and reference for project portfolio selection in enterprises.

作为战略性的管理工具，项目组合选择强调“做正确的项目”。以往的项目组合选择研究多假设决策者为“完全理性”，较少考虑决策者的行为因素，导致做出决策的参考价值有限。本课题采用模糊集理论刻画项目参数，综合“前景理论”和“有限理性”理论，提出模糊环境下基于参照依赖的项目组合选择模型。具体的研究内容包括：(1)分析单参照点、多参照点的确定方式和它们对决策者行为的影响，构建满意度函数；提出模糊环境下项目组合选择风险度量的CVaR方法。(2) 现实约束下，考虑决策者的“模糊厌恶”行为，建立了满意度最大、项目组合收益的模糊性和含糊性最小的多目标项目组合选择模型。(3) 根据模型的特点，设计了一种混合算法用于提出模型的求解；(4)将提出的模型和算法应用于新药R&D项目，与传统的模型进行对比；考察不同参照点对决策结果的影响。本课题拓宽了经典项目组合选择的理论，研究成果将为企业的项目组合选择提供参考和帮助。

本课题的主要研究成果如下：1）将行为决策理论融入到项目组合选择问题，考虑决策者的偏好行为，将行为决策领域较新的TODIM方法应用于能源项目管理。2）研究了新能源项目组合选择的评价指标体系，分析了传统的TODIM方法在新能源项目评价的适用性，给出了一种多粒度语言变量的改进TODIM方法，并进一步探析了决策者偏好行为的影响。3）综合目前对合同能源管理商业模式的不足，本课题对合同能源管理商业模式的影响机制进行了全面的梳理，并给出了一种新的TODIM方法应用于合同能源管理商业模式的选择方法。4）针对项目组合选择涉及的财务因素和非财务因素，构建了一个两阶段的模糊项目组合选择模型，通过建立风险补偿率与非财务风险值的函数，构建了考虑非财务风险的项目组合选择模型。5）项目组合选择模型是一个非线性和具有复杂约束的组合优化问题。传统的优化方法对于这类问题往往难以求解。智能优化算法在求解这类问题时表现出良好的性能。本课题针对项目组合选择问题难于求解的特点，提出了几种高效的求解智能优化算法。在研究成果的基础上，本课题在SSCI或SCI检索的国际期刊发表正刊文章15篇，其中项目负责人以第一作者或通讯作者发表JCR Q1学术论文11篇。另有一些论文正在审稿中，还有一批论文即将投稿。