A multi-objective optimization framework integrating ICSL deep learning for forecasting and scheduling emergency medical supply demand in public health emergencies.

To address the prediction and allocation challenges of emergency medical supplies during the middle and late stages of major outbreaks, we proposed an end-to-end ICSL deep learning algorithm architecture. This approach takes into account the characteristics of infectious diseases and the impact of government quarantine measures on their spread, enabling the prediction of the maximum demand for emergency medical supplies. Based on this, a multi-objective scheduling and allocation model was constructed, considering urgency, scheduling time, and cost. We designed a multi-objective particle swarm optimization algorithm to solve this model. Finally, we validated the algorithm using data from the Wuhan pandemic control measures. The results showed that the parameter update method improved the prediction accuracy of the LSTM model, increasing accuracy by 29.37% compared to the traditional LSTM algorithm and by 8.63% compared to the improved BP neural network algorithm. The proposed scheduling and allocation model optimizes delivery timeliness while also considering the urgency and cost-effectiveness of the resource allocation. The research findings can provide decision-making insights for the allocation of emergency medical supplies during public health emergencies.