The use of heuristic ordering and particle swarm optimization for nurse scheduling problem
The scheduling of nurse has been the important management functions and directly affected the hospital services and the patient care as well. The challenge of nurse scheduling is the substantial amount of time for assigning shifts to a set of nurses that involves a large set of constraints. To satis...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2017
|
| Subjects: | |
| Online Access: | http://eprints.utem.edu.my/id/eprint/20625/1/The%20Use%20Of%20Heuristic%20Ordering%20And%20Particle%20Swarm%20Optimization%20For%20Nurse%20Scheduling%20Problem.pdf http://eprints.utem.edu.my/id/eprint/20625/2/The%20use%20of%20heuristic%20ordering%20and%20particle%20swarm%20optimization%20for%20nurse%20scheduling.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The scheduling of nurse has been the important management functions and directly affected the hospital services and the patient care as well. The challenge of nurse scheduling is the substantial amount of time for assigning shifts to a set of nurses that involves a large set of constraints. To satisfy all the constraints cannot be applied directly due to variability preferences and conflicting interest and objectives between nurses and hospital. The main objective of this research study is to find the optimal solution that can deal with varying preferences in term of skill categories, flexible constraint parameters, flexible coverage and varying the size of nurses. A combination of heuristic ordering and particle swarm optimization (HOPSO) has been proposed to achieve the objectives. The capability of PSO algorithm is enhanced by emphasizing the use of information on the constraints and heuristic ordering for searching optimal solution in both the feasible and infeasible solution spaces. The constraints are adapted to the evaluation function that iteratively evaluates all the solutions. The solution with lowest violation penalty cost is selected and will compare to a new solution. Before the particles update its position, variable neighborhood search (VNS) is applied in order to enhanced the diversity before being trapped in a local optima. The performance of the proposed method is tested on the real problem benchmark dataset in Malaysia public hospital, Universti Kebangsaan Malaysia Medical Centre (UKMMC) that had 8 datasets with the different number of nurses. The comparison of the result of HOPSO, harmony search algorithm (HSA) and heuristic variable neighborhood search (HVNS) is presented. The result show that the proposed HOPSO can generate the schedule in the range between one to twenty-six seconds computational time, followed by the HSA which range between one hundred and eighty-five to three hundred and forty-five seconds and HVNS takes one hundred and thirty-one to eight hundred and twenty seconds. HOPSO can decrease the penalty cost into ninety seven percent improvement than the HSA which is less than fifty percent of improvement. Computational results show that the proposed HOPSO based algorithm is performed better than HSA and HVNS in order to provide a practical solution to the problem. |
|---|