Overcrowding in emergency departments (ED) is manifest through excessive patients waiting time. Excessive waiting time has been linked to both health concerns and increased stress for employees. In order to alleviate overcrowding in emergency departments, a correct utilization of available resources is crucial. To this end, human resource scheduling has been the topic of many research papers in emergency departments. However, to capture the detailed structure of the system, most studies focus on an approach combining heuristics and simulation. This paper proposes an optimization scheduling model that strategically allocates human resources in the ED. The optimization model is based on a queuing system that captures the inner system dynamics of the ED. In addition, multiple resources are considered and scheduling is done simultaneously to account for their inherent interactions. Consequently, given a fixed staffing level, the model finds the optimal combination of shifts so as to minimize the total patients waiting time. In order to capture the uncertainties in emergency departments, the scheduling model uses non-stationary stochastic arrival rates and stochastic service times. The system's parameters are defined using admissions data from the university hospital center in Lille (France). The primary aim of the study is to provide EDs with a robust cyclic schedule to be used under varying conditions. To this end, different scheduling hypotheses are tested and evaluated against the feasibility of the resulting schedules. The solutions are then validated using a simulation model that replicates the structure of the optimization model and serves to introduce further complexity.