【作者】 赵小华；

【导师】 杨育；

【作者基本信息】 重庆大学 ， 机械工程， 2011， 博士

【摘要】 随着生产规模的不断扩大以及经营的国际化,国内生产企业以大规模投资为特征的R&D项目成为了中国企业进入世界500强的重要途径,项目组合模式在此过程中扮演了极为重要的角色。这类企业通常都是快速成长型生产企业,其R&D项目组合在组织结构、项目主体、项目控制等方面包含了大量柔性因素。目前生产实践中采用的传统项目管理技术与方法,由于过多强调操作层面并依靠确定信息的控制与优化,导致了较多项目未能实现预定目标。本文在系统研究国内外相关成果的基础上,以提升成长型生产企业R&D项目组合的控制水平为目标,针对快速成长型的生产企业R&D项目组合动态管理进行了深入研究,主要内容包括:①构建了成长型生产企业R&D项目组合动态管理概念模型。在对项目管理基础理论、技术、工具进行研究的基础上,分析了动态管理概念模型构建的必要性和应遵循的原则,进而针对成长型生产企业R&D项目组合管理的特点与要求,构建了以柔性因素为核心、全周期、立体化的R&D项目组合动态管理概念模型;概念模型从项目周期和管理层级两个维度提出了项目组合动态管理的四个核心问题,即项目组合优化配置(What)、智力资源筛选(Who)、矩阵组织研究(How)和项目组合优化调度仿真(Who&When&What)等,为后续研究奠定了基础。②研究了成长型生产企业R&D项目组合优化配置的权衡模型及其求解算法。项目组合优化配置是项目组合管理的重要战略决策,针对决策指标分散导致模型过于复杂及其柔性因素难以评价两大问题,结合国内成长型生产企业注重技术能力成长的特点,论文确立了技术能力、收益满意度和交叉相似度等三个决策指标,并分别使用技术能力成长模型、定性概率理论和模糊评价等面向柔性技术实现其量化,进而建立R&D项目组合优化配置模型,设计了求解算法,最后通过一个应用实例说明了该模型及算法的有效性与实用性。③研究了成长型生产企业R&D项目组合智力资源筛选模型及其求解算法。在成长型生产企业R&D项目组合智力资源匹配决策中,为减少候选人能力、机会成本与协同效率等因素带来的不确定性与模糊性,引入能力成熟度模型思想,提出一种基于三角模糊数与分级算法的候选人能力成熟度评价方法,较好地解决了候选人能力难以定量评价的问题。在5级模糊评价候选人机会成本和协同效率的基础上,建立业主组织成员选择的能力-成本-协同效率的多目标权衡模型,使用遗传模拟退火算法进行求解,为智力资源筛选提供方法与决策依据,通过应用案例阐明该方法的应用过程,验证其合理性与有效性。④研究了成长型生产企业R&D项目组合矩阵组织柔性系统动力学模型。组织结构及其管理策略对成长型生产企业R&D项目组合成败有重要影响。为更深入理解矩阵式结构的构成及其相互依赖关系,发掘其动态反馈机制,论文结合成长型生产企业R&D项目特点,提出基于子域与全域的两级系统动力学的组织建模方法,并重点通过一个应用实例详细论述其建模过程,最后基于反馈循环及干预点提出了针对案例组织的改进策略。⑤研究了成长型生产企业R&D项目组合优化调度模型及其仿真。针对成长型生产企业R&D项目组合共享资源冲突等因素导致的组合进度难以估计问题,论文分析了成长型生产企业R&D项目组合调度的柔性特点,构建了基于市场机制的多Agent仿真模型,通过循环竞标解决共享资源动态匹配问题;为体现能力对项目进度的动态影响,构建了基于能力演化模型及效率转化函数的资源占用时长动态模型;为模拟子项目竞标过程,以资源成本和拖期惩罚为目标构造子项目竞标策略效用函数,并使用改进遗传模拟退火算法进行求解。最后,通过一个仿真实例验证了该方法的可行性与有效性。更多还原

【Abstract】 Along with the technology innovation and the enterprise expansion, the R&D projects have been an important approach for domestic production enterprises to enter Top 500 list. The project portfolio management plays an important part in this process. The R&D project portfolios in production enterprises are featured of being full of flexible factors in many aspects, such as the organization structure, the projects, the process of project controlling et al. The traditional project management tools focus on the operational level and heavily depend on“hard”information. This fact eventually leads to many failures in completing project goals. Based on the domestic and abroad researches on this topic, a dynamic management method considering flexible factors for R&D project portfolios are studied. The main contents are fivefold:①The conceptual model of dynamic management method considering flexible factors is constructed. Based on the review of the fundamental theories and the popular tools of project management, the necessity of dynamic management model and the principles for constructing one are analyzed. Combining with the characteristics of R&D project portfolio in production enterprises, a full-cycled, full-leveled model centered on flexible factors is built, in which, four key control process of portfolio dynamic management are systematically addressed from dimensions of both project cycle and management level.②The tradeoff model in R&D project portfolios and its solving algorithm are studied. Portfolio selection is one of the most important strategic decisions in R&D project portfolio management. However, the decision indices for the decision process are uncertain and difficult to evaluate. To address these problems, the capability of technology, the satisfactory degree of project return and the crossing similarity are chosen as the decision indices. To quantitively evaluate those indices the model of evolution of capability of technology, the qualitative possibility theory and the fuzzy evaluation method are introduced respectively. The tradeoff model of portfolio selection for R&D project portfolios is then constructed and the algorithm to solve the model is developed. The feasibility and effectiveness of the model and solving algorithm is verified by a case study.③The model for member selection and its solving algorithm are studied. To reduce the uncertainty and ambiguity in the member selection process of R&D project portfolios, the influence of candidates’competence, opportunity cost and coordination efficiency is investigated. By introducing the CMM (Capability Maturity Model) concept, a capability maturity evaluation method is presented based on triangular fuzzy numbers and a ranking formulation. Then a 5-graded rating method is proposed for the assessment of opportunity cost and coordination efficiency. To find the optimal member selection solution, a competence-cost- coordination efficiency trade-off model is then developed, and SAGA (Simulated annealing genetic algorithm) method is proposed to solve it. The model is applied and its rationality and effectiveness is verified.④The soft system dynamic model for matrix structure project organization is studied. The organization structure and the corresponding management strategies are much coherent to the outcome of the project. To better understand the organization structure and interdependencies, a Composite Soft System Dynamic Modeling method (CSSDM) is developed. The composite modeling method based on sub-domain and whole-domain models is proposed. The case of a R&D project management organization modeling is elaborately analyzed to show the process. Based on the causal loops and the leverage points in model output, the proposition of strategy improvement is presented.⑤The simulation model for project portfolio scheduling was studied. To address the difficulty in R&D project portfolio scheduling, a simulation model based on market mechanism is developed. An iterated bidding process is adopted to dynamically allocate shared human resources. To reflect individual influences on scheduling, a dynamic resource occupation model is developed with the capability evolution model and the efficiency mapping function. To simulate the bidding process, a utility function of agents’bidding policies is constructed with objectives of resource cost and delay penalty. A case is studied to verify its feasibility and effectiveness.更多还原