Weighted Tabu Search for Multi-Stage Nurse Rostering Problem

Zhouxing SU,Zhuo WANG,Zhipeng L(U)

DOI: https://doi.org/10.1360/n112015-00284

2016-01-01

Scientia Sinica Informationis

Abstract:This paper presents a study of the multi-stage nurse rostering problem, which was proposed in the Second International Nurse Rostering Competition.We propose a weighted tabu search algorithm for solving this rostering problem, which is a very important problem in healthcare optimization.The algorithm employs the following neighborhood structures: 3 structures that are exclusive to each other and 1 compound structure.The possibilities of selecting a neighborhood to search are tuned dynamically by the extent to which they are adaptive.Meanwhile, the balance between intensification and diversification during the search procedure is achieved by adjusting the weight of the penalty for each nurse.We address the problem caused by the lack of global information within each independent stage by proposing an approximate evaluation strategy for spanning constraints.The algorithm employs a cache for neighborhood move evaluation with the purpose of reducing the overall computational cost for the adaptive neighborhood selection strategy, which improves the efficiency of the algorithm further.The benchmark result of our algorithm on 60 instances from the competition shows good solving quality and ranks fourth in the final of the Second International Nurse Rostering Competition.Furthermore, we compare and analyze some key factors in the algorithm and demonstrate their rationality.