فهرست نشریات علمی دسترسی آزاد ایران

This study presents a novel deep learning–based multi-objective optimization framework for minimizing financial risks in smart supply chain finance. The proposed model integrates deep neural networks for dynamic credit risk prediction with evolutionary optimization algorithms such as NSGA-II to simultaneously minimize risk exposure, reduce capital costs, and enhance liquidity stability. Using synthetic and real financial data, the framework captures complex nonlinear patterns in supply chain interactions and translates them into adaptive decision-making strategies. Comparative analysis against baseline models demonstrates superior predictive accuracy, broader Pareto front coverage, and higher robustness under market fluctuations. Sensitivity analysis further confirms the model’s resilience to changes in key financial parameters such as interest rates, credit limits, and payment delays. The results highlight the potential of combining deep learning and multi-objective optimization to enable data-driven, risk-aware, and sustainable financial decision-making in digital supply chains.