Introducing a novel multi-objective optimization model for volunteer assignment in the post-disaster phase: Combining fuzzy inference systems with NSGA-II and NRGA
Metadatos
Mostrar el registro completo del ítemEditorial
Elsevier
Materia
Volunteer assignment Humanitarian Decision support system Multi-objective optimization Evolutionary algorithms Fuzzy inference system (FIS)
Fecha
2023-04-15Referencia bibliográfica
P. Rabiei et al. Introducing a novel multi-objective optimization model for volunteer assignment in the post-disaster phase: Combining fuzzy inference systems with NSGA-II and NRGA. Expert Systems With Applications 226 (2023) 120142[https://doi.org/10.1016/j.eswa.2023.120142]
Patrocinador
H2020-EU.1.3. – EXCELLENTResumen
Each year, disasters (natural or man-made) cause a lot of damage and take many people’s lives. In this situation,
many volunteers come to help. While the proper management of volunteers is very effective in controlling the
crisis, the lack of proper management of volunteers can create another crisis. Therefore, we introduce a model to
deal with the volunteer assignment problem by considering two qualitative objective functions: The first one is
minimizing the mean importance of Emergency Department (ED) centers’ unmet needs by volunteers, and the
second one is minimizing the mean degree of unsatisfied preferences of selected volunteers. To evaluate the
introduced qualitative indexes, two Fuzzy Inference Systems (FISs) are used to encapsulate decision makers’
knowledge as well as the human reasoning process. FISs are embedded in two evolutionary algorithms for solving
the proposed model: Non-Dominated Sorting Genetic Algorithm II (NSGA-II) and Non-Dominated Ranked Genetic
Algorithm (NRGA). Also, 30 small-size problems, as well as 30 large-size problems, are randomly generated
and solved by both metaheuristic algorithms. Using the obtained data, the performance of NSGA-II and NRGA is
measured and compared based on four criteria: CPU Time, Number of Non-dominated Solutions (NNS), Mean
Ideal Distance (MID), and Spacing Metric (SM). Statistical tests show that both algorithms have the same performance
in small-size problems. However, in large-size problems, NSGA-II is faster, and NRGA produces more
optimal solutions. The proposed model is flexible enough to adapt to different scenarios just by updating linguistic
rules in FISs. Also, since employed algorithms produce a set of optimal solutions, decision-makers can
easily choose the most appropriate solution among the Pareto front based on the circumstances