时间缓冲设置与鲁棒性项目调度

发布时间：2018-06-05 05:07

本文选题：鲁棒性项目调度 + 质量鲁棒性　；参考：《华中科技大学》2014年博士论文

【摘要】：由于市场环境快速变化,项目特别是大型工程项目所涉及的不确定性因素日益增多,使越来越多的项目出现问题,如费用超支、超出工期甚至中途夭折等。与此同时,现代项目管理又要求周期更短、准时完工率更高、成本更低。在这种情况下,应对复杂环境中出现的各种不确定因素,保证项目按计划按时完工而形成的鲁棒性项目调度,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。时间缓冲技术是建立鲁棒性项目调度计划的主要方法,主要有关键链集中缓冲方法和分散缓冲方法。对这两种缓冲方法进行研究,建立合理的鲁棒性计划,成为该领域的重点问题。本文综合应用双目标优化、启发式算法及系统仿真等方法,对二种缓冲的适用环境及其相应的鲁棒性项目调度计划进行了系统深入的研究,主要研究工作如下：首先,综合考虑项目执行环境中不确性因素,如活动时间确定性程度、项目交货期限,按时完工重要性等,分别运用集中缓冲和STC分散缓冲方法,制定鲁棒性项目调度计划,在仿真环境下执行该计划,对比二种缓冲方法的鲁棒性绩效指标。结果表明：分散缓冲法具有更好的解鲁棒性；当活动时间的不确定性程度较大时,集中缓冲法的质量鲁棒性较好；当项目工期较紧的情况下,分散缓冲法具有更好的鲁棒性。然后,针对不同的项目网络特征,研究二种缓冲方法的适用性。综合考虑项目规模、网络复杂度等因素,生成不同的项目实例,并利用两种缓冲方法分别制定鲁棒性项目调度计划,模拟不确定环境下项目计划的执行。仿真结果表明：当活动时间不确定程度较小时,在小规模简单的项目中,分散缓冲鲁棒性优于集中缓冲,这时优先选择分散缓冲法制定鲁棒性项目调度计划。当项目规模较大且网络复杂时,或在项目不确性程度较大时,集中缓冲的质量鲁棒性始终优于分散缓冲,这时应优先选择集中缓冲法。接着研究了双目标鲁棒性项目调度算法。为使项目既保证按时完工又能按计划执行,同时考虑解鲁棒性和质量鲁棒性,并以STC分散缓冲为基础,构建了双目标鲁棒性调度模型。综合运用模拟退火和禁忌搜索算法,提出了求解这种双目标鲁棒性调度问题的二阶段算法。实验结果表明,本文所提出的算法搜索过程较快稳定性较好,所得解优于单智能算法解,能够根据管理者的鲁棒性偏好获得双目标调度问题的近优解。最后,对非关键链溢出和关键链断裂情况下的关键链计划进行研究。针对关键链计划插入接驳缓冲后出现资源再次冲突、关键链断裂、非关键链溢出等问题,提出两种解决方案。第一种方案从缓冲设置方式的视角进行研究,用分散缓冲方法替代集中缓冲,最后通过项目管理数据库(PSPLIB)中大量实例,证明该方法的普遍适用性及其在完工和计划鲁棒性方面的优势。第二种方案用二阶段启发式算法对出现问题的关键链计划直接进行重排：第一阶段运用动态规划和三种重排优先级规则对计划进行重排来消除资源冲突。针对重排后仍然存在关键链断裂或非关键链溢出的项目,运用第二阶段提出的启发式算法,进行二次重排,得到合理的关键链计划。最后,经过单项目和大量项目仿真实验证实,该方案能够有效解决大部分项目的关键链重排存在的问题,且所制定的关键链进度计划在仿真执行中具有更好的完工性绩效。
[Abstract]:Due to the rapid change of the market environment , the uncertainty factors involved in the project , especially the large - scale project , have caused more and more project problems , such as overspending of the cost , higher on - time completion rate and lower cost . At the same time , the modern project management requires shorter cycle , higher on - time completion rate and lower cost .

Firstly , the uncertain factors of the project execution environment , such as the degree of certainty of the activity time , the time limit of the project , the importance of time completion , etc . are comprehensively considered , and the robust project scheduling plan is formulated by using the centralized buffer and STC dispersion buffer method respectively , and the robustness performance indexes of the two buffering methods are compared under the simulation environment .
when the degree of uncertainty of the activity time is large , the quality robustness of the concentrated buffer method is better ;
The decentralized buffer method has better robustness when the project duration is tight .

Then , the applicability of two kinds of buffering methods is studied for different project network characteristics . Considering the factors such as project size and network complexity , different project instances are generated , and the implementation of the project plan under the environment is simulated . The simulation results show that , when the project size is large and the network is complicated , or when the project is not true , the quality robustness of the centralized buffer is always better than the decentralized buffer , and then the central buffering method should be preferred .

A two - stage algorithm for solving the problem of double - target robust scheduling is proposed by using simulated annealing and tabu search algorithm . The results show that the proposed algorithm is better in stability and better than single intelligent algorithm , and can obtain the near optimal solution of double - target scheduling problem according to the robustness preference of the manager .

In the end , the key chain plans in the case of non - critical chain overflow and critical chain fault are studied . Two solutions are proposed for the key chain plans . First , the key chain plans of the problem are rearranged by using dynamic programming and three rearrangement priority rules . Finally , through single project and large number of project simulation experiments , this scheme can effectively solve the problem of key chain rearrangement in most projects , and the key chain schedule plan has better completion performance in simulation execution .
【学位授予单位】：华中科技大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TU71

【参考文献】