Developing two heuristic algorithms with metaheuristic algorithms to improve solutions of optimization problems with soft and hard constraints: An application to nurse rostering problems

Abstract

Many researchers have studied optimization problems with soft and hard constraints, such as school timetabling, nurse rostering, vehicle routing with soft time window, and job/machine scheduling. Nurse rostering problem (NRP) is the research problem in this paper. This study proposes two heuristic algorithms, which are the decision tree method and the greedy search algorithm, to integrate with metaheuristic algorithms in order to generate better initial solutions in less time and to improve solutions’ quality. This research examines the algorithms’ performance based on two scenarios and two metaheuristic algorithms: bat algorithm (BA) and particle swarm optimization (PSO). For the two scenarios, BA (or PSO) with the decision tree method outperforms BA (or PSO) without the decision tree method, and BA (or PSO) with the greedy search algorithm outperforms BA (or PSO) without the greedy search algorithm. Furthermore, the results show that BA (or PSO) with the decision tree method and the greedy search algorithm can generate better initial solutions in less time and improve solutions’ quality.