Abstract. This paper introduces an extension to the well-established Resource-Constrained Project Scheduling Problem for the comprehensive description of disruption management problems. This conceptual framework employs the concept of alternative activities to consider both the temporal shift of activities or the reallocation of resources and switches from one valid process variant to another one. Activities can be serialized or parallelized, process steps can be inserted or removed and durations as well as resource requirements can be modified dynamically during optimization. Focusing on the domain of the aircraft turnaround as the most important airport ground process, we illustrate how the Extended RCPSP (x-RCPSP) can be applied for decision support. A specific evolutionary algorithm is presented that identifies good-quality solutions to relatively large disruption management problems within only a few seconds. The results of the evaluation illustrate fast convergence on good or optimal schedules and serve as a basis for the development of future problem solving algorithms.

Abstract. This paper describes how the Resource-Constrained Project Scheduling Problem (RCPSP) can be used as a basis for real-time decision support in the disruption management of the aircraft turnaround, the most typical airport ground process. For this purpose, the RCPSP is extended by the possibility to describe alternative activities, which can be used to model potential process modi cations. An evolutionary approach is presented for solving this generalized problem, considering both basic rescheduling possibilities as well as the option of exchanging activities. The work presented in this paper is based on the results of a study conducted in collaboration with Deutsche Lufthansa AG, concerned with the analysis of the elementary requirements of decision support systems for turnaround process management.

Experiments show that LRS outperforms previous approaches. †

No abstract

Abstract-Whenever an unforeseen disturbance occurs during the execution of scheduled operations, rescheduling might be necessary: Beside temporal shifts and the allocation of alternative resources, also potential switches from one process variant to another one have typically to be considered. In realistic scenarios of operational disruption management (DM) the high number of potential options makes the provision of online decision support complex. It is thus necessary to significantly reduce the size of the regarded (search) problems which can for instance be achieved by applying methods of partial rescheduling. However, existing approaches such as Affected Operations Rescheduling (AOR) or Matchup Scheduling (MUP) focus on production-specific problems and can not be applied to more generic problem classes. To overcome this limitation, we introduce a novel approach to partial rescheduling in this paper: Local Rescheduling (LRS) is based on the incremental extension of a time window which is regarded for potential schedule modifications. We discuss how this time window can be initialized, extended and used for rescheduling. Moreover, we illustrate the superior performance of LRS in comparison with full rescheduling and MUP.