Meta-heuristics for solving nurse scheduling problem: A comparative study

Nurse Scheduling Problem (NSP) is a very day-to-day problem categorized as a hard problem due to several constraints associated with it. Some of the constraints are from the employer's or hospital's point of view and some are from the employees' or nurses' point of view. An optimized solution is required violating as minimum constraints as possible. While the problem has been solved by many nature-inspired algorithms like Simulated Annealing (SA) and Genetic Algorithm (GA), we have also implemented these algorithms in addition to taking on board another two comparatively newer meta-heuristic methods - namely Firefly Algorithm(FA) and Particle Swarm Optimization (PSO). After implementing the four of them, we compared their performance mainly based on optimality. We found out that Firefly Algorithm gave best results both in terms of number of solved instances and final cost of the solution. However, Particle Swarm Optimization algorithm showed poor performance compared to the other algorithms and gave the least number of feasible solutions.