Proactive-reactive procedures for robust project scheduling

The vast majority of research efforts in project scheduling concentrates on developing procedures to generate workable baseline schedules that minimize the project makespan in a deterministic environment. However, a real-life project is inevitably subject to uncertainty during execution. This dissertation aims at introducing procedures that cope with disruptions during execution. We limit ourselves to the treatment of time uncertainties caused by the fact that actually realized activity durations during project execution may deviate from the expected activity durations. When dealing with uncertainty in a scheduling environment, there are, generally spoken, two main approaches. Proactive scheduling focuses at incorporating safety in the schedule to absorb future disruptions, while reactive scheduling denotes how to react when a disruption occurs. Both approaches are inescapably related. We will investigate in how several proactive-reactive scheduling decisions can help a project manager to increase the quality of a project. The text of this dissertation is organized as follows. Chapter 1 introduces the problem of proactive-reactive project scheduling and situates it in the extensive project scheduling literature. Chapter 2 formulates the problem at hand and defines the concepts required in the remainder of the thesis. The trade-off between makespan and stability in project scheduling of Chapter 3 justifies the research efforts made in Chapters 4, 5 and 6 to add safety to the baseline schedule. Mainly two approaches to add safety are discussed in this thesis. First, a schedule is made proactive in Chapter 4 by deciding how the resources flow throughout the project. Next in Chapter 5, we develop efficient and effective procedures to add idle time (buffers) into a schedule to anticipate unforeseen events. Chapter 6 contains an extension of Chapters 4 and 5 by merging scheduling, resource allocation and buffer allocation decisions into an integrated approach. The reactive procedures that are required to decide how to react on disruptions that cannot be absorbed by the baseline schedule are introduced in Chapter 7. In Chapter 8 an extensive simulation-based experiment is set-up to evaluate several predictive-reactive resource-constrained project scheduling procedures proposed in the previous chapters. Chapter 9 applies the proactive-reactive project scheduling methodology on a real-life project that stems from our experience in the Belgian construction industry. Accordingly, a risk management framework is introduced to detect and analyze the risks that constitute the uncertainty implied in the project. In a last chapter, some overall conclusions and recommendations are provided.status: publishe