Blood supply chain location-inventory problem considering incentive programs: comparison and analysis of NSGA-II, NRGA and electromagnetic algorithms

• Published in: Neural computing & applications 2024-11, Vol.36 (31), p.19469-19487
• Main Authors: Alikhani, Tayebeh, Dezfoulian, Hamidreza, Samouei, Parvaneh
• Format: Article
• Language: English
• Subjects: Artificial Intelligence; Blood; Blood & organ donations; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Science; Cost analysis; Data Mining and Knowledge Discovery; Genetic algorithms; Heuristic methods; Image Processing and Computer Vision; Indicators; Original Article; Probability and Statistics in Computer Science; Sorting algorithms; Supply chains
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-024-10216-z

Problem Blood is a rare perishable substance with limited life in the real world and blood supply chain management is a vital subject. Hence, it is trying to design an efficient supply chain network to create a balance between blood supply and demand, particularly in deficient conditions. One effective solution for blood deficiency compensation is the use of incentive programs at the right times to encourage people for blood donation. The novel aspect of this study considers a new mathematical model to design a blood supply chain network with the location of temporary centers for collecting donated blood, in addition to incentive programs in the right periods to actualize the goal of creating blood supply-demand equilibrium and minimizing the cost of the network. Method In this paper, four methods have been used in different dimensions to solve the proposed model. In this case, augmented epsilon constraint (AEC/EC) was used for small dimensions, while electromagnetic algorithm (EM), Non-dominated ranked genetic algorithm (NRGA), and non-dominated sorting genetic algorithm (NSGA-II) were used for large dimensions due to the inherent complexity of the problem. Results The performance of algorithms was analyzed based on the four standard indicators. Then, their outputs were evaluated using statistical assumption tests at the significance level of 0.05. In three considered indicators (SNS, MID, and TIME indicators), the NSGA-II algorithm outperformed the NRGA and EM algorithms. This case indicated the superiority of the NSGA-II algorithm over the NRGA and EM algorithms especially for problem solution time, which is one of the most significant indicators used in metaheuristic algorithms.