deferential evolution algorithm

The employee scheduling seeks to find an optimal schedule for employees according to the amount of demand (workload), employee availability, labor law, employment contracts, etc. The importance of this problem in improving the quality of service, health and satisfaction of employees and reducing costs, including in hospitals, military or service centers, has encouraged researchers to study. In this regard, nurse rostering problem is a scheduling that determines the number of nurses required with different skills and the time of their services on the planning horizon. In this research, by adding the nurses' shift preferences and number of consecutive working days constraints, an attempt has been made to make the problem more realistic. The objective function of the problem is to minimize the total cost of allocating work shifts to nurses, the cost of the number of nurses required to reserve, the cost of overtime from a particular shift, the cost of underemployment from a particular shift, the cost of overtime on the planning horizon, the cost of underemployment on the planning horizon and the cost of absence shift-working and non-working days preferred by nurses. To solve problem, after modeling the problem as a mixed-nteger program and due to the complexity of the problem, the differential evolutionary algorithm is used with innovation in its crossover operator. To validate the proposed algorithm, its output was compared with the genetic algorithm. The results show that the differential evolutionary algorithm has good performance in problem-solving. Keywords: Nurse Rostering Problem, Deferential Evolution Algorithm

The importance of resource leveling cannot be ignored throughout a project, especially in cases where a trivial fluctuation in using resources will lead into high financial risks or costs. Traditional approaches are usually insufficient in solving the problems of multi-objective constraint resource leveling. In this paper, we propose an optimized technique based on Differential Evolution (DE) algorithm and ELECTERE method in order to overcome the drawbacks of the traditional resource leveling algorithms. Furthermore, obviating the challenges of normalizing and weighting objective functions, this approach takes the uncertainty and ambiguity of the objective functions into account through application of pseudo-criterion. The structure of proposed model is designed based on real situations. Results obtained from the numerical example verify the efficiency of the proposed model in solving the problems of resource leveling, and in comparison with traditional models; they confirm the better performance of this model to overcome single and multi-objective problems.