|Ph.D Student||Sellam Balouka Noemie|
|Subject||The Multi-Mode Resource-Constrained Project|
Scheduling Problem with Activity Values and
Uncertain Activity Durations
|Department||Department of Industrial Engineering and Management||Supervisor||Dr. Izack Cohen|
|Full Thesis text|
This work addresses the multi-mode resource-constrained project scheduling problem (MRCPSP). This problem involves several components of project management, such as resource constraints and a trade-off between time and cost. In the current work, we propose two significant extensions of this typical project management problem. \par First, we formulate a model that combines quantitative project management models that focus on time and cost with value-focused qualitative methodologies by extending the MRCPSP to include value aspects (e.g., technical performance). Indeed, the current trend in modern project management methodologies is to maximize project values. We suggest a quantitative model that balances time, cost and value. It generates detailed project plans, which are feasible with respect to time, cost and resource constraints, and considers interactions between technical performance, time and cost. Such interactions, which existing project management models typically ignore, or project managers deal with intuitively, are important for increasing projects' values. We develop a specialized solution approach that solves the problem to near-optimality in reasonable computational times. Also, the model is able to generate an efficiency frontier of project plans, where each plan on the frontier achieves the best value for its cost and it enables decision makers to select their preferred plan.
Secondly, we use a robust optimization approach for making tactical project decisions in uncertain environments. Stochasticity is manifested by uncertain activity durations, which leads to stochastic resource demands and costs. The objective is to minimize the project duration while taking uncertainty into account. To the best of our knowledge, we formulate the first robust counterpart of the MRCPSP, when durations are varying within a polyhedral uncertainty set; we develop an analytical solution approach, and examine its performance compared to other alternatives. We discuss how the level of conservatism and other conditions affect the price of robustness. Finally, we propose heuristic procedures for solving the robust counterpart for large projects, and we compare their performances.