Über die Entwicklung von Crew Scheduling Systemen in der Schifffahrt
Vortrag gehalten beim KoSSE-Tag 2019 "Software Engineering für Logistik 4.0", 23. September 2019, Lübeck
The increased capacities of merchant ships in the last decades as well as tighter port frequencies are the main reasons for a higher workload onboard ships. The size and composition of ship crews, the levels of their qualification and training are the main factors for securing their ability to control and handle situations in a safe way. Furthermore, a lack of qualified crew scheduling ability in daily ship operation may result in a lack of workload control in combination with fatigue issues, which in turn can lead to major accidents. In this work the problem of determining optimal crew demand by calculating work schedules under consideration of different legislative and safety requirements is formally specified and a model as well as solution methods for solving real-world data instances are presented. The problem is classified as a scheduling one and belongs to the class of NP-hard problems from a complexity theoretical point of view. A task-based approach is used to specify the main inputs as the workload, seafarer's qualifications and the voyage specific dependencies. Company specific requirements as well as complex legislative regulations for rest hours are modeled based on their importance as hard or soft constraints in order to guarantee a good and robust mapping of the real world problem. Heuristic methods with smart decomposition of the feasible region and problem specific sorting methods are implemented and used for the problem solution. Solution methods are designed by performance and enable fast calculation of schedules directly on board giving the ship commands the possibility to react to changes in voyage with appropriate updated work schedules, avoid fatigue and guarantee a safe ship operation. This paper contains a detailed description of how the decision support system (DSS) is used to analyze scenarios and discusses several criteria, which can be applied in the evaluation of sufficient crew. The model is parameterized in such a way that different strategic decisions with immediate impact on seafarer's well-being e.g. watch keeping patterns can be varied and tested easily. The DSS has been validated for the merchant ship domain and could be used for further safety-critical systems in maritime domain.