YI-CHANG LIHUNG-CHANG LIAO2026-10-162026-10-162012-01-01Pure ID: 15213101Pure UUID: 852ba5da-2942-4d85-b92b-73ecf8a53b1fhttps://www.scopus.com/pages/publications/84869787990https://scholars.csmu.edu.tw/handle/123456789/30906The objective of this study is to design a robust blood supply chain system considering the total cost (TC) and the safety of patients. The freshness of transfused blood affects patient safety. The turnover rate (TR) is used in this study as the index of a patient's safety level. This study examines the blood supply chain system of the Taichung Blood Center of the Taiwan Blood Services Foundation and its contracting hospitals. A robustly designed blood supply chain system considers the input, the optimal approach, and the output. The input stage of the design uses the dynamic Taguchi method to set the signal, noise, and control factors of the blood supply chain system. The approach of this study was to use a neural network (NN) and a genetic algorithm (GA) to determine the results of the output process. The results show that the most robust control combination is a FIFO (first in, first out) blood delivery policy (BDP) where the maximal inventory level (MaxIL) is 600, the minimal inventory level (MinIL) is 80, and the donor arrival rate (DAR) is 120. Finally, the sensitivity analysis is discussed.16enjournal article