Megow, NicoleMöhring, Rolf H.Schulz, Jens2021-12-172021-12-1720092197-8085https://depositonce.tu-berlin.de/handle/11303/15670http://dx.doi.org/10.14279/depositonce-14443This paper concerns the highly complex task of scheduling large-scale maintenance activities during a plants shutdown or turnaround. We model it as a discrete time-cost tradeoff problem with capacity constraints and individual working shifts for different resource types with a cost function regarding a balanced resource consumption. We introduce and model the problem, give an overview on the large variety of related optimization problems, and propose a framework for supporting managers decisions in the planning process of such an event. Our key component is an optimization algorithm complemented with a risk analysis of solutions. We implemented a two-phase solution method in which we first provide an approximation of the tradeoff between project duration and cost as well as a stochastic evaluation of the risk for meeting the makespan. In a second, detailed planning phase, we solve the actual scheduling optimization problem for a chosen deadline heuristically and compute a detailed schedule that we complement by evaluating upper bounds for the two risk measures expected tardiness and the probability of meeting the deadline. We present experimental results showing that our methods can handle large real-world instances within seconds and yield a leveled resource consumption. For smaller instances a comparison with solutions of a time-consuming mixed integer program prove the high quality of the solutions that our fast heuristic produces.en510 Mathematikschedulingmalleable jobstime-cost tradeoffproject managementstochastic analysisoptimizationresource levelingResearch Paper