基于缓冲管理的项目进度控制方法研究

发布时间：2018-03-03 14:21

本文选题：关键链缓冲管理　切入点：项目控制　出处：《华中科技大学》2016年博士论文　论文类型：学位论文

【摘要】：当前市场环境快速变化且竞争愈发激烈,项目所面临的不确定性和风险不断增大,同时项目本身的结构和规模口趋复杂,对稀缺资源的竞争逐渐加剧,导致越来越多的项目难以按期完工,传统的项目管理方法比如关键链路径法、计划评审技术等,已经越来越难以应对现代项目面临的高要求和复杂多变的执行环境。关键链缓冲管理方法CC/BM应运而生,迅速成为项目管理领域理论研究的热点,同时在实践中比如生产、制造及服务等行业得到了成功应用。缓冲是其中的重要概念,通过插入集中的时间缓冲以构建资源可行的鲁棒调度计划,旨在吸收不确定性并聚合风险；同时,通过缓冲监控来实施进度控制,以保证项目按期完工。因此,缓冲的大小估计和监控方法,直接决定着项目计划完工期及项目进度风险,是关键链缓冲管理技术运用成功的关键。本文综合应用风险管理理论、项口进度管理理论以及鲁棒性调度相关理论,采用贝叶斯网络、资源流网络、Monte Carlo模拟、启发式算法等方法,对基于缓冲管理的项目进度控制方法展开系统深入的研究,主要工作概述如下：首先,提出了考虑活动工期风险和多资源约束风险的缓冲管理方法。该方法运用贝叶斯网络技术分析关键风险因素、评估其发生概率及对活动工期的影响；通过资源流网络评估资源受限程度,结合活动资源需求强度提出了资源风险系数计算方式。综合以上两种风险,提出合理的缓冲大小计算方法,并设计了一个动态执行环境下基于贝叶斯网络风险溯源、风险传递和项目资源流网络更新的动态缓冲监控模型,以有效应对项目进度风险。通过算例分析验证了所提缓冲大小计算方法的实用性和有效性。然后,针对现有的缓冲监控机制忽视相关的活动信息、未能充分考虑项目执行的动态特征这一问题,将一种基于活动的进度风险分析法与缓冲监控方法结合起来。即将前者中的活动关联度指标引入后者,作为有效触发赶工行动的决策依据,从而构建了一种全方位(项目层-活动层)的项目进度控制框架。此外,提出两种关联度监控阈值的动态设置方法,即根据项目进展比例或者缓冲消耗水平动态调整行动阈值。模拟实验证明了集成进度控制方法相对于传统缓冲监控方法的优越性,尤其是采用递减方式设置关联度监控阈值时效果最好。接着,针对现有缓冲监控方法在指导赶工决策时都没有考虑赶工成本信息这一不足,提出了考虑活动赶工成本的项目进度控制过程。为了更好地描述基于关键链计划的赶工问题,我们首先建立了一个数学优化模型。考虑到数学优化方法的应用局限性,继而提出一种基于Monte Carlo模拟的启发式求解方法,通过衡量项目按时完工率的改进与活动所需赶工成本之比,并结合当前缓冲消耗水平,确定赶工候选活动集合以及当前决策点活动的具体赶工时间长度。仿真实验结果显示了该方法相对于相对缓冲监控法在项目时间及成本绩效方面的优越性。最后,从资源可用量约束和反应式调度的角度研究了基于缓冲的项目进度控制问题,考虑到资源成本和进度稳定性提出了一个两阶段进度-成本控制系统。在计划阶段,基于期望资源成本采用模拟优化方法计算常规资源分配截止时点；在执行阶段当缓冲消耗超出监控阂值时,选取具有最大总成本节约的活动,对其增加紧急资源以使其提前开始,以此修复项目进度并保持计划稳定。通过算例分析阐明所提方法的应用步骤,并设计大规模模拟实验对影响各种方法绩效的实验参数进行了分析。
[Abstract]:The rapid change of the market environment and the increasingly fierce competition, the project faced with uncertainty and risk is increasing, at the same time, the structure and size of the project itself and more complex, competition for scarce resources gradually increased, resulting in more and more projects to be completed on schedule, traditional project management methods such as key chain path method, program evaluation and review technique that has become increasingly difficult to meet the high demands of modern project facing the complicated and changeable execution environment. CC/BM critical chain buffer management method came into being, quickly become a hot research field of project management theory, and in practice, such as the production, manufacturing and service industries have been successfully applied. The buffer is one of the important concepts, by inserting the focus the buffer time to construct robust scheduling resource plan feasible, to absorb uncertainty and aggregate risk; at the same time, through buffer monitoring The implementation of schedule control, to ensure that the project be completed. Therefore, the size estimation and monitoring method of buffer, directly determines the project risk and project completion schedule, critical chain buffer management techniques is the key to success. Based on the theory of risk management, project schedule management theory and robust scheduling theory, using Bayesian network, cyber source flow, Monte Carlo simulation, heuristic algorithms, the research of the system of schedule control methods of buffer management based on the project, the main work is summarized as follows: firstly, the buffer management method considering activity duration risk and resource constrained risk. By using the method of key risk factors of Bayesian network technology. To assess the impact of the occurrence probability and the durations of activities; through the resource flow network assessment resource constrained degree, combined with active resource demand The strength of computing resources. The risk coefficient of the above two kinds of risk, proposes the calculation method of buffer size is reasonable, based on Bayesian network risk investigation and design a dynamic execution environment, dynamic buffer monitoring model of risk transfer and project resource flow network update, to effectively deal with the project schedule risk. Through the example analysis shows the practicability and validity of the proposed buffer size calculation method. Then, according to the neglect of the existing mechanism of buffer monitoring activities related to information, failed to give full consideration to the problem of the dynamic characteristics of the execution of the project, a schedule risk analysis method based on activity and buffer monitoring methods together. The association index is the former activity the introduction of the latter, as the trigger action rush decision basis, so as to construct a full range (project layer and active layer) project schedule control box Frame. In addition, the dynamic setting method of two kinds of correlation threshold monitoring, according to the progress of the project or the proportion of buffer consumption level dynamic adjustment action threshold. Simulation experiments show that the integration schedule control method is superior to the traditional buffer monitoring method, especially by decreasing set correlation monitoring threshold was the best. Then, according to the the existing buffer monitoring method in guiding decisions are not considered rush rush this lack of cost information, considering the activities crashing project control process. In order to better describe the problem of critical chain scheduling based on time, we first establish a mathematical optimization model. Considering the application limitation of mathematical optimization method, and then put forward a heuristic method based on Monte Carlo simulation, and through improving the activities needed to catch the measure of project completion rate Engineering cost ratio, combined with the current buffer consumption level, determine the candidate set and the current work activities of the activities of the specific decision points work length of time. The simulation results show that this method is superior to the corresponding buffer monitoring method in the project time and cost performance. Finally, from the resource availability constraints and reactive scheduling perspective study on the problem of the project schedule control buffer based on considering the resource cost and schedule stability proposed a two stage of progress cost control system. In the planning stage, the conventional calculation method of simulation and optimization of resource allocation based on resource cost using the expected cut-off point; in the implementation stage when the buffer consumption beyond the monitoring threshold, it selects the maximum total cost saving activities, increase emergency resource to the early start, in order to repair and maintain the stability of project schedule plan. Through the example The application steps of the proposed method are analyzed, and the experimental parameters that affect the performance of various methods are analyzed in the design of large-scale simulation experiments.

【学位授予单位】：华中科技大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：F273

【相似文献】