【作者】 杨颖；

【导师】 杨善林；

【作者基本信息】 合肥工业大学 ， 管理科学与工程， 2011， 博士

【摘要】 随着科学技术的不断进步，产品的功能和结构日益复杂，制造业的竞争越来越体现为复杂新产品在经济和科技实力上的竞争。许多企业更是把复杂新产品的开发作为重要的利润源泉和获得持续性竞争优势的关键。为充分利用有限的研发资源，企业往往会针对不同目标市场的客户需求，同时设计开发多种复杂产品。多个并行的复杂产品开发项目之间相互争夺资源，为企业级的项目管理带来困难，也对相关理论研究提出了迫切要求。目前的研究成果大多侧重于单个项目的管理和资源配置，对企业级的项目组合管理研究较少，且多数针对普通的研发项目。因此面向复杂产品开发项目，研究企业级项目管理中的动态决策问题，从企业战略的高度对复杂产品开发项目组合进行协调管理，对于提高复杂产品开发的效率和资源利用率具有重要的理论意义和实践意义。本文针对复杂产品开发企业多项目管理的现状，系统研究了复杂产品开发项目组合的动态决策问题，提出了基于企业战略的动态决策模式和相关决策方法。主要内容包括建立项目全生命周期的评价指标，对复杂产品开发项目与企业战略的一致性进行评价；探讨项目之间的相互依赖性，构建定性分析和定量分析模型对依赖关系和依赖程度进行描述；设计出基于企业战略的项目组合动态决策模式，并综合考虑项目与企业战略的一致性及项目间的相互依赖关系构建复杂产品开发项目组合的动态决策模型；最后为整个项目组合决策过程构建了综合集成型的决策支持系统。本文的主要研究内容和创新点如下：（1）研究了复杂产品开发项目与战略的动态匹配问题。首先分析复杂产品开发项目全生命周期各阶段的特征，基于平衡计分卡理论从价值与目标、客户与市场、流程与管理、学习与成长等四个维度构建项目的战略性评价指标；然后考虑指标间的因果关系，采用网络分析法确定各指标的权重，建立了BSC-ANP评价模型；同时，考虑评价过程中信息的不确定性，采用证据理论对多层指标上的定性和定量信息进行合成，并将证据理论与平衡计分卡相结合形成了基于BSC-ER的评价方法；在某轿车公司的整车平台开发项目与战略的一致性评价中进行实例应用，验证了该评价方法的合理性，可以有效解决项目与战略的匹配问题。（2）研究了复杂产品开发项目间相互依赖关系及依赖程度的识别问题。首先从企业采取的经济模式和产品开发过程两方面分析项目组合产生的原因，深入探讨复杂产品开发项目间依赖关系的产生机理；然后分析项目的生态学属性，并基于生态学属性建立了包含收益依赖、成本依赖、进度依赖和风险依赖等四个维度的定性分析模型；最后对项目之间的依赖性进行定量分析，分别采用依赖矩阵及关联性分配矩阵方法建立了两个项目之间及多个项目之间的依赖性分析模型，定量地描述了各种依赖存在的情况下对项目目标值的影响。（3）研究了复杂产品开发项目组合的动态决策问题。首先分析企业战略对复杂产品开发的影响机理，建立了企业战略与复杂产品开发项目组合决策过程的相互影响模型；根据项目组合决策应遵循的原则，设计了基于企业战略的项目组合动态决策模式。然后，在该决策模式的指导下，综合考虑项目与企业战略的一致性及项目间相互依赖关系，构建了项目组合动态决策的非线性整数规划模型，采用改进的模拟退火算法对该模型进行求解。在某企业的轿车整车开发项目组合决策中的实例应用验证了模型和算法的合理性与有效性，并且为企业级的项目管理提供了战略性指导框架。（4）研究了复杂产品开发项目组合的动态决策支持系统。针对项目组合决策模型的复杂性，分析和构建了一个决策支持系统来辅助整个组合管理过程。首先分析复杂产品开发项目组合决策系统的特征，给出了基于综合集成的项目组合决策支持方法；然后设计了人机交互的综合集成型决策支持系统，并给出了系统的体系架构和功能组成，讨论了系统实现的关键技术；最后以某轿车整车开发企业为背景，结合其多个复杂车型并行开发的特点，应用上述理论研究成果，为其构建了一个综合集成型决策支持系统。更多还原

【Abstract】 With the continuous improvement in science and technology, the function and structure ofproducts become more and more complex. Competition among manufactures reflects the economicand technological strength of complex new product development. Complex product development isregarded increasingly as an important source of profit and a key element to access the sustainablecompetitive advantage. Many companies often design and develop a variety of complex productssynchronously to satisfy customer needs from different target markets, which can make full use ofthe limited R&D resources. More and more complex product development projects compete forresources. So it is urgent to study how to manage the synchronized multiple projects. Currentresearch focused on selection and resource allocation of a single project mostly and only on R&Dprojects. There are fewer literatures on project portfolio management in enterprise level. Therefore,it is necessary to study dynamic decision making problems in complex product development (CPD)project portfolio management. Managing and coordinating complex product development projectportfolio has theoretical and practical significance. It can improve resource utilization and increasethe efficiency of product development.The dissertation focused on dynamic decision making problems of complex productdevelopment project portfolio considering the current status of multi-project management inenterprises. It put forward a dynamic decision making model and respective methods based oncorporate strategy. Firstly, evaluation index system of projects was developed and project’sconsistency with strategy was measured in the life cycle; Then, interdependence between projectswas explored and formulated with qualitative and quantitative analysis model; Thirdly, a dynamicproject portfolio decision-making pattern was designed and dynamic decision-making methods werepresented, which considered the project’s consistency with strategy and interdependences betweenprojects. Finally, a meta-synthesis decision support system was built to assist the whole decisionmaking process.The main research contents and contributions in the dissertation are summarized as follows:(1) Research on the problem of matching between CPD projects and corporate strategy. Firstly,characteristics of each stage in the life cycle of CPD process was analyzed and an evaluation indexsystem was developed based on Balanced Scorecard(BSC) to measure the project’s alignment fit with corporate strategy. The system included four aspects: values and objectives, customer andmarkets, process and management, learning and growth. Then BSC-ANP assessment model wasestablished, in which the causal relationship between indicators was considered and the weight ofeach index was determined by the Analysis Network Process (ANP) method. Meanwhile, because itis full of uncertainty information in the evaluation process, evidence reasoning(ER) was introducedto synthesize qualitative and quantitative information. A evaluation method based on BSC-ER wasproposed which intergrated BSC with ER. Lastly, a case study in an automobile company showedthat the evaluation method is reasonable. It can resolve the problem of matching between vehicleplatform development projects and corporate strategy.(2) Research on the interdependence identification among CPD projects. Firstly, the causes ofthe project portfolio were explored from two aspects. One is the economic model adopted bycorporate; the other is the product development process. Then the mechanism of interdependenceamong projects was analyzed. The interdependence was described from objective properties andecological properties of projects. The qualitative analysis model was developed, including fivedimensions: benefit dependence, cost dependence, schedule dependence, risk dependence and logicdependence. Lastly, quantitative analysis on interdependencies among CPD projects was carried out.The interdependence between two projects and the interdependence among multiple projects wereformulated respectively. These models described the effect of the presence of dependencies on theproject target value quantitatively.(3) Research on the problem of dynamic decision-making of CPD project portfolio. Firstly, theinfluence of corporate strategy on the project portfolio decision-making process was analyzed and aproject portfolio dynamic decision-making pattern was established based on strategy. In this pattern,CPD project portfolio decision-making process was divided into four interrelated sub-processes.They were portfolio strategic positioning process, evaluating the project’s consistency with strategyprocess, interdependencies analyzing process, dynamic portfolio selecting process. Then, a nonlinearinteger programming model on dynamic portfolio decision-making was presented under theguidance of that pattern, considering the project’s consistency with strategy and interdependenciesamong projects. A improved simulated annealing algorithm was designed to solve the model. Therationality and effectiveness of the model and the algorithm were verified by a case study in anautomobile company.(4) Research on dynamic decision support system of the CPD project portfolio. Because theportfolio decision-making process was complex and the decision-making model was difficult toapply in practice, a decision support system was built to assist the whole decision process. Firstly,characteristics of CPD project portfolio decision-making system was analyzed and a solution to complex decision problems based on meta-synthesis methods was presented. Then, a comprehensiveintegrated interactive decision support system was designed. The system architecture and functionalcomponents were presented and the key technologies of implementation were discussed. Finally,research results above were combined with an example of a case company, in which there weremany concurrent complex product development projects. A meta-synthesis decision support systemwas built to aid the company’s portfolio decision process.更多还原