การจัดเส้นทางการจัดส่งโลหิตด้วยวิธีวิวัฒนาการโดยใช้ผลต่างแบบผสมผสาน กรณีศึกษากลุ่มโรงพยาบาลในจังหวัดสงขลา

บทคัดย่อ

This research has studied the blood delivery for the 12th Regional Blood Centre (RBC), Songkhla province, where is the center of blood distribution to hospitals in Songkhla province and other provinces in lower South of Thailand. The blood delivery of this RBC was modelled as the capacitated vehicle routing problem under the known demand of blood. This work has proposed the differential evolution hybridized with the new local search, called DE+NLS algorithm, to find out the optimal route with shortest total distance. For the testing of performance, it was executed on the 30 datasets of real-world case study that were collected during the period of August, 2021. The performance of DE+NLS, in addition, were compared with the current routing method, standard differential evolution (DE) algorithm and simulated annealing (SA) algorithm. As a results, the DE+NLS produced the route having less total distance than the DE and SA. The overall total distance obtained by the DE+NLS was reduced from 10,804.20 kilometers (by current routing method) to 4,017.38 kilometers. The statistical results clearly indicated that the DE+NLS was an efficient method and it significantly outperformed the DE and SA methods. The DE+NLS spent less time than the SA while it consumed slightly more than the DE due to the local search procedure. Moreover, the DE+NLS achieved the lowest transportation cost at 17,463.54 Baht, which was decreased by 61.82% from the current method. Thus, the DE+NLS was a well-design and efficient algorithm for the given problems.