摘要
传统的资源受限项目调度问题属于确定性项目调度问题，假定项目具有完全信
息且项目执行环境静态不变。然而，这样的假定在现实中通常不成立，因为定义一
个项目的所有参数在项目执行阶段都可能发生改变。不确定因素使得基于确定性假
设产生的调度方案难以有效实现既定的调度目标。反应性项目调度是不确定性项目
调度的一个重要分支，指的是：在项目执行阶段，干扰事件发生后，基准调度方案
变得不再可行或最优时，对基准调度方案进行修改或者再优化的过程。在许多情况
下，为了规避风险和提升项目绩效，保证新调度方案的最优性仍然是反应性项目调
度需优先考虑的问题。但是，既有文献中，尚无这方面的研究。既有的反应性项目
调度方法忽略了确定性项目调度目标的最优实现，因而无法保证反应性调度方案在
新的项目执行环境下是最优的。
鉴于此，本文提出一个考虑资源可用量不确定性的多模式项目反应性调度问题
和一个考虑新项目到达的多项目反应性调度问题。与既有的反应性项目调度问题不
同，本文提出的反应性项目调度问题，先保证新的调度方案在新的项目执行环境下
是最优的，再最小化新旧项目调度方案之间的偏差。同时，基于 CPLEX V12.8.0，
提出一种两阶段反应性项目调度方法。在第一阶段，对项目进行完全重调度，得到
确定性项目调度目标在新的项目执行环境下可实现的最优值；在第二阶段，以最优
实现确定性项目调度目标为约束，以最小化基准调度方案的调整成本为目标，求得
一个新的最优调度方案作为反应性调度方案。最后，基于标准算例，对本文提出的
两阶段反应性调度方法与既有的反应性调度方法（含完全重调度方法）进行充分的
比较测试。
结果表明，本文提出的两阶段反应性调度方法能兼顾基准调度方案的调整成本
与确定性项目调度目标的最优实现，与既有方法相比，具有明显的优势。这将有助
于丰富反应性项目调度领域的相关理论，并满足项目管理者在项目管理实践中的特
定需求。
关键词：资源受限项目调度；反应性调度；多项目调度；不确定性项目调度
I

Abstract
The traditional resource-constrained project scheduling problem is a deterministic
project scheduling problem, with the assumption that the project has complete information
and the project execution environment is static. However, this assumption is usually not
tenable in reality because all parameters defining a project may change during the process
of project implementation. Uncertain factors make it difficult for the schedule based on the
deterministic assumption to effectively achieve the established scheduling objectives.
Reactive project scheduling refers to the process of modifying or re-optimizing the baseline
schedule when it becomes no longer feasible or optimal after the occurrence of disruption
events during the process of project implementation. In many cases, in order to avoid risks
and improve project performance, ensuring the optimality of the new generated schedule is
still a priority for reactive project scheduling. However, there is no research in this field
among the existing literature. Reactive project scheduling is an important branch of project
scheduling under uncertainty. The existing reactive project scheduling methods ignore the
optimal realization of the deterministic project scheduling goal, so they cannot ensure that
the reactive schedule is optimal in the new project execution environment.
To address this issue, this thesis proposes a resource-constrained multi-mode project
reactive scheduling problem with resource availability uncertainties and a resource-
constrained multi-project reactive scheduling problem with new project arrival. Different
from the existing reactive project scheduling problems, the reactive project scheduling
problems proposed in this thesis first ensure that the new generated schedule is optimal in
the new project execution environment, and then minimize the deviation between the old
schedule and the new one. In addition, a two-stage reactive project scheduling method based
on CPLEX V12.8.0 is proposed. In the first stage, the optimal value of the deterministic
project scheduling goal in the new project execution environment can be obtained by fully
rescheduling the project. In the second stage, a new optimal schedule achieving the new
optimal value of the deterministic project scheduling objective is formulated to minimize
II

the adjustment cost of the baseline schedule. Finally, based on the benchmark instances, the
proposed two-stage reactive scheduling method and the existing two methods, including the
full rescheduling method, are sufficiently compared and tested.
The results demonstrate that the proposed method not only considers the adjustment
cost of the baseline schedule, but also optimally achieves the deterministic project
scheduling objective, with distinct advantages over the existing methods. This will help to
enrich the relevant theories in the field of reactive project scheduling and meet the specific
needs of project managers in project management practice.
Keywords: resource-constrained project scheduling problem; reactive scheduling; multi-
project scheduling; project scheduling under uncertainty
III
。。。以下略