【作者】 别黎；

【导师】 崔南方；

【作者基本信息】 华中科技大学 ， 管理科学与工程， 2012， 博士

【摘要】 经济全球化导致市场竞争日趋激烈,现代项目日趋复杂,要求周期更短、准时完工率更高、成本更低,这对项目管理而言提出了更高的挑战。作为一种新的项目管理方法,关键链项目管理在实践中特别是在不确定性较大的项目环境中得到了成功应用,引起了企业界和学术界的广泛关注,成为项目管理领域理论研究的热点。本文综合应用项目计划与控制理论、约束理论,采用活动相关性模型和贝叶斯模型,对关键链项目管理中的缓冲估计与监控方法进行系统深入的研究,其中主要的研究工作如下：首先,针对现有单项目环境中的缓冲大小估计方法假设独立活动工期的不足,提出了活动相关条件下的缓冲大小计算法。该法在对活动工期相关性进行建模基础上,提出了四种基于活动相关性的参数,并通过大量计算实验去分析这四种参数对项目工期均值和标准差的影响,将其中影响明显的参数,即相关程度和相关因子,加入到缓冲大小计算公式中。通过算例分析表明所提方法在项目活动相关程度明显或者不确定性程度较大时,会相应地产生较大缓冲,提供更好的保护。其次,针对多项目环境下的能力约束缓冲大小估计问题,提出了基于“鼓”活动特点的新方法。该法分析使用“鼓”资源的活动与单项目中接驳链的不同之处,综合分析影响能力约束缓冲大小的因素：“鼓”资源紧张度、“鼓”网络复杂度和“鼓”活动间隙度。在确定这三个因素的度量方法后,将其加入到缓冲大小计算公式中,从而得到更加准确的缓冲大小,以减弱不确定因素的影响和改善多项目管理绩效。通过算例分析验证了该法的实用性和优越性。此外,针对项目执行的动态环境和不确定程度逐渐减少的情况,提出了动态缓冲监控基本模型。此模型根据项目执行情况,来动态计算缓冲大小,动态设定监控点,动态调整两个触发点的高低,从而发出与实际项目进展相符的预警信息。接着对基本模型进行扩展,即利用线性贝叶斯模型重新估计后面活动的工期,并采用活动相关性自适应程序来更新缓冲大小。通过算例分析验证了该法的实用性和优越性。最后,从项目进度监控的角度研究了动态缓冲监控问题。针对现有缓冲监控方法由于忽视项目内部结构以及活动信息而产生不准确预警信号的不足,将缓冲监控与进度风险分析法结合起来。引入其中的活动敏感性信息,将其作为区别不同活动的决策依据。当缓冲消耗量到达黄区时加入基于活动CRI的监控过程。在动态设置CRI的监控阀值基础上,提出了基于活动敏感性信息的关键链动态缓冲监控模型。实验结果表明,合理设置CRI监控阀值后,所提混合监控方法的综合监控绩效更优。更多还原

【Abstract】 With rapid development of economic and market, the environment of project becomes more and more complicated, resulting from shorter due date, higher on-time completion rate and lower cost. Project management is facing a greater challenge. As a new project management methodology, Critical Chain Project Management (CCPM) has been successfully applied in the project environment with large uncertainty. Many theoretical books and articles have already appeared on the subject, and it also received considerable attention among practitioners, so CCPM has become one of the most hot and popular topic in the field of project management. Based on the synthetical application of project planning and control methods, Theory of Constraints, and statistical dependence model and Bayesian methods, this dissertation makes a systematic study of buffer sizing method and buffer monitoring method. The main research works are as follows:Firstly, the assumption of independence in most existing buffer sizing approaches is unrealistic, and a method for determining buffer sizes with dependence assumption between activities is introduced. In detail, the dependence relationship between activities is represented based on the model in schedule risk analysis method, and four dependence parameters are proposed. The effect of these parameters on project duration performance is analyzed by means of a large number of computational experiments. Two parameters with the obvious impacts, which are the dependence degree and the dependence factor, are integrated into the formulation of buffer sizing approach. The suggested method is tested and compared to the methods with independence assumption. The results indicate that the suggested method can provide better protection when at least one of the two definitions is at a high level.Secondly, a new method to determine the capacity constraints buffer size was proposed based on the characteristics of drum activities in critical chain multi-project scheduling. In detail, the difference between drum activities and activities on the feeding chain were discussed. Effects of drum resource tightness, drum network complexity and drum space interval were measured and integrated in this method, so that the various uncertain problems are solved and the project performances are improved in multi-project management. A simulation experiment was carried out to demonstrate the effectiveness and practicability of the proposed method.Thirdly, based on the decrease in the degree of uncertainty under dynamic environment, a dynamic buffer monitoring method was proposed. This method dynamically calculates the buffer size and the time instant for monitoring, adjusts the place of the two control trigger points to control the project, so that the proper warning information can be sent to the project manager to take correct action to meet project deadline. Then, the process of updating activities duration was illustrated by using the Linear Bayesian Method to expand the basic buffer management model. A simulation experiment was carried out to demonstrate the effectiveness and practicability of the proposed method.Finally, the dynamic buffer monitoring problem is studied from the perspective of project schedule monitoring. The current buffer monitoring methods may generate incorrect warning information because they neglect the project structure and activity informations. To solve this defect, the buffer monitoring method is combined with the schedule risk analysis method. Specifically, the activity sensitivity information is introduced as a basis for decision making. An activity cruciality index based bottom-up project monitoring method is triggered and integrated into buffer control process when buffer consumption penetrates into the yellow region. Experimental results show that the proposed hybrid monitoring method do better than the current buffer monitoring methods in the monitoring performance indicators when we reasonably set the action threshold of CRI.更多还原