An Experimental Application of Multi-Objective Evolutionary Algorithm to Many-Objective Nurse Scheduling for Real General Hospitals

This paper describes an attempt of nurse scheduling by means of many-objective evolutionary algorithm (MaOEA). The nurse schedule is supposed to be used by a nurse organization in an actual general hospital. MOEA/D is adopted as MaOEA for optimization. By adding some restrictions to initialization and mating operations for the schedule, the constraints of the optimization problem are invalidated, and the problem is formulated as an unconstrained combinatorial optimization problem. Other requirements are formulated as objectives of the problem. The simplified MOEA/D optimizes the schedule.