复杂工程多项目中专享资源规划与鲁棒调度集成优化

Multi-project management is a major pattern of commercial operation in today’s highly competitive marketplace. Concurrent projects, especially complex engineering projects, present themselves to be more widely distributed, more specialized and team-oriented. Besides, the execution of modern projects are increasingly faced with various uncertainties, demanding high robustness of project schedules. All these factors impose a big challenge to traditional multi-project scheduling approaches that are based on a resource sharing policy, since it is difficult to perform efficient resource investment or to produce stable execution plans. This research is conducted from the perspective of a resource dedication policy, in which the resources dedicated to one project are not allowed to be shared with other parallel projects. The integrated optimization of resource planning and robust scheduling in a multi-project environment subject to various uncertainties is studied, where multiple objectives, i.e. tardiness cost, resource cost and robustness measures, are simultaneously considered. More specifically, we first construct a chanced-constrained programming model of proactive scheduling by modeling the uncertainty of resource availabilities in multi-project organizations. The solution methods based on the uncertain programming theory are investigated. Then we propose a two-stage proactive approach to cope with activity duration uncertainty by combining multi-project resource planning with the time buffering technique. Finally, a multi-objective optimization model of reactive scheduling that allows emergency resource allocation is established, and a rolling horizon based local rescheduling algorithm is designed for solving the model. The research has important practical significance for multi-project organizations to deal with real-world planning and scheduling problems by the successful application of robust project scheduling methods.

在激烈的市场竞争中，多项目管理成为现代商业运营的主要模式。并行执行的多个项目特别是复杂工程项目，越来越呈现出分布地域广、专业化和团队化的特征以及面临的不确定因素多、对鲁棒性要求高等特点，传统的基于资源共享策略的多项目调度方法难以满足对有限资源有效投入、进度计划稳定执行等需求。本课题基于资源专享视角，针对多项目执行中的各种不确定因素，在考虑拖期成本-资源成本-鲁棒性的多目标问题框架下，研究多项目整体规划中的资源投入和鲁棒调度集成优化方法。通过对多项目中资源总量不确定性建模，提出资源不确定条件下的前摄式调度机会约束规划模型及求解算法；将多项目资源规划与时间缓冲相结合，研究工期不确定条件下的两阶段前摄式鲁棒调度方法；考虑紧急资源调配策略建立反应式调度多目标优化模型，设计基于滚动窗口的局部重调度算法。研究成果对指导多项目组织成功应用鲁棒项目调度方法解决实际中的规划与调度问题，具有重要的现实意义。

在激烈的市场竞争中，多项目管理成为现代商业运营的主要模式。并行执行的多个项目特别是复杂工程项目，越来越呈现出资源竞争激烈、面临的不确定因素多、对鲁棒性要求高等特点，传统的确定性（多）项目调度方法难以满足对有限资源有效分配、进度计划稳定执行等需求。本课题提出一类不确定环境下的多项目资源分配与鲁棒调度集成优化问题，从“前摄调度——反应调度”两个层次开展研究。首先，针对活动工期的不确定性，提出了基于鲁棒资源分配的多项目前摄调度方法，通过定义基于资源的二元决策变量来表示资源传递路径，以最小化资源转移成本和最大化项目鲁棒性为目标建立数学规划模型，分别设计了遗传退火混合智能算法、非支配排序遗传算法和Pareto模拟退火算法对双目标优化模型求解。其次，同时考虑多项目拖期成本和鲁棒性目标，提出基于时间缓冲的多项目前摄调度方法，构建了一个资源专享-转移视角下的多项目资源投入（战术层）与鲁棒调度（运作层）双层决策优化模型，设计了一种多目标自适应大邻域搜索算法求解模型。进一步地，将鲁棒资源分配与时间缓冲策略相结合，研究了工期不确定条件下的两阶段前摄式鲁棒多项目调度方法，目标为最小化 “总工期-鲁棒性成本”双目标的加权和。接着，针对关键链多项目反应调度问题，分别提出四种进度计划重排策略（包括并行/串行调度、鲁棒并行/鲁棒串行调度）以及两种行动策略（包括活动赶工、紧急资源调配），并设计大规模仿真实验，分析了应用不同的关键链识别优先级指标、设置不同的缓冲尺寸以及采取不同的反应策略对关键链多项目鲁棒性能的交互影响。最后，在传统资源受限多项目调度问题中引入资源转移时间，分别在确定和活动工期不确定条件下研究了带资源转移时间的多项目（鲁棒）调度模型和启发式优化算法。研究成果对指导多项目组织成功应用鲁棒项目调度方法解决实际中的规划与调度问题，具有重要的现实意义。