فهرست نشریات علمی دسترسی آزاد ایران

This research presents an intelligent and integrated framework that combines deep learning-based forecasting and multi-objective optimization to improve decision-making in humanitarian supply chains during crises significantly. The designed model can process incomplete and noisy data, extract hidden patterns in demand behavior and crisis severity, and dynamically incorporate this information into the resource allocation process, consistent with real-world conditions. The results show that this framework simultaneously improves forecast accuracy, response speed, and distribution fairness, and also has stable performance in severe uncertainty scenarios. Scenario-based analyses and comparisons with baseline methods show that the proposed model can provide more optimal operational options and increase the resilience of the relief network. These achievements indicate that the proposed framework provides a new path for the development of intelligent systems in crisis management.