Data-Driven Dynamic Adjustment and Optimization Model of Emergency Logistics Network in Public Health

BACKGROUND AND AIM: In the long-term prevention of the COVID-19 pandemic, parameters may change frequently for various reasons, such as the emergence of mutant strains and changes in government policies. These changes will affect the efficiency of the current emergency logistics network. Public health emergencies have typical unstructured characteristics such as blurred transmission boundaries and dynamic time-varying scenarios, thus requiring continuous adjustment of emergency logistics network to adapt to the actual situation and make a better rescue. PRACTICAL SIGNIFICANCE: The infectivity of public health emergencies has shown a tendency that it first increased and then decreased in the initial decision-making cycle, and finally reached the lowest point in a certain decision-making cycle. This suggests that the number of patients will peak at some point in the cycle, after which the public health emergency will then be brought under control and be resolved. Therefore, in the design of emergency logistics network, the infectious ability of public health emergencies should be fully considered (ie, the prediction of the number of susceptible population should be based on the real-time change of the infectious ability of public health emergencies), so as to make the emergency logistics network more reasonable. METHODS: In this paper, we build a data-driven dynamic adjustment and optimization model for the decision-making framework with an innovative emergency logistics network in this paper. The proposed model divides the response time to emergency into several consecutive decision-making cycles, and each of them contains four repetitive steps: (1) analysis of public health emergency transmission; (2) design of emergency logistics network; (3) data collection and processing; (4) adjustment and update of parameters. RESULTS: The result of the experiment shows that dynamic adjustment and update of parameters help to improve the accuracy of describing the evolution of public health emergency transmission. The model successively transforms the public health emergency response into the co-evolution of data learning and optimal allocation of resources. CONCLUSION: Based on the above results, it is concluded that the model we designed in this paper can provide multiple real-time and effective suggestions for policy adjustment in public health emergency management. When responding to other emergencies, our model can offer helpful decision-making references.