Deep Reinforcement Learning Model for Blood Bank Vehicle Routing Multi-Objective Optimization

The overall healthcare system has been prioritized within development top lists worldwide. Since many national populations are aging, combined with the availability of sophisticated medical treatments, expenditures rapidly growing. Blood banks a major component any system, which store and provide blood products needed for organ transplants, emergency routine surgeries. Timely delivery is vital, especially in settings. Hence, process parameters such as safety speed have received attention literature, well other cost. In this paper, time cost modeled mathematically marked objective functions requiring simultaneous optimization. A solution proposed based on Deep Reinforcement Learning (DRL) to address formulated Multi-objective Optimization Problems (MOPs). basic concept decompose MOP into scalar optimization sub-problems set, where each one these separate Neural Network (NN). model sub-problem optimized neighborhood parameter transfer DRL training algorithm. step undertaken collaboratively optimize model. Pareto-optimal solutions can be directly obtained using trained NN. Specifically, multi-objective bank problem addressed research. One technical advantage approach that once available, it scaled without need retraining. scoring straightforward computation NN layers limited time. technique provides set strength points ability generalize solve compared methods. was tested 5 hospitals Saudi Arabia’s Riyadh region, simulation results indicated decreased by 35% 30%, respectively. particular, outperformed state-of-the-art Genetic Algorithms Simulated Annealing.