Seminar: Krzysztof Postek (TU Delft)

Real-life project or machine scheduling involves: (i) limited information about the exact task durations, and (ii) an opportunity to reschedule each time a task completed its processing and a machine becomes idle. Robust optimization is the natural methodology to cope with the first characteristic, yet the existing literature does not consider the possibility to adjust decisions as more information about the tasks' durations is revealed. This is despite that re-optimizing the schedule is a standard practice.

We develop an approach that takes into account, at the beginning of the planning horizon, the possibility that scheduling decisions can be adjusted, allowing an arbitrary set of scenarios for the task lengths' realizations. We demonstrate that this can lead to better here-and-now decisions. In a recent work, we develop an exact MILP formulation of this problem for discrete sets of scenarios, which, to best of our knowledge, is the first one in which so-called nonanticipativity constraints are formulated exactly.

Joint work with Shimrit Shtern (Technion), Izack Cohen (Bar Ilan University), Izak de Heer (TU Delft).