A greedy constructive approach for Nurse Rostering Problem

Abstract

Nurse Rostering Problem (NRP) concerns about producing a high quality workable duty roster for the available staff nurses. The aim of this work is to present a greedy constructive heuristic algorithm to generate a feasible initial solution by satisfying the hard constraints. Basically the initial solution includes three steps: first we start by designing a group of shift patterns based on hard and soft constraints. Then, those patterns are rotated for predefined positions and allocated to each nurse. Finally; if the solution is not feasible we use a repair mechanism. In this work, a real world problem from Universiti Kebangsaan Malaysia Medical Centre (UKMMC) is used to test the proposed algorithm. The resulting roster demonstrates that our proposed algorithm generates a good quality duty roster in a reasonable computational time for our case study.