Optimizing employee schedules by a hybrid genetic algorithm

Creating an employee schedule means taking into account many heavy constraints like employee contracts or minimal staffing levels on the one hand and many global, difficult to formalize constraints like aspects of fairness on the other hand. Optimisation is quite difficult especially when fix rostering schemata cannot be used, e.g. because of frequently varying staffing levels. In this paper we present how real-life employee scheduling problems can be solved by applying a Hybrid Genetic Algorithm that uses problem specific knowledge. First we briefly describe the given problem domain, then the idea and implementation of the Genetic Algorithm is presented. Finally we show some application results and the outlook.