A moment-based Optimization Model for Designing the Supply Chain of Dairy Products: Data-driven and Sustainable Approach
Mahyar Abbasian, Amin Jamili
Abstract:
This study presents a moment-based optimization model for designing a sustainable, data-driven supply chain for perishable dairy products. The proposed multi-objective model integrates economic, environmental, and social dimensions of sustainability and addresses the inherent uncertainty in demand through a machine learning forecasting approach. The supply chain network includes producers, distributors, and retailers, with the aim of minimizing total costs and carbon emissions while maximizing job creation. A novel moment-based reformulation is introduced to enhance computational tractability, allowing the model to be efficiently solved using state-of-the-art optimizers such as Gurobi. Additionally, a CNN-based algorithm is employed for route optimization and fitness evaluation, improving decision-making under dynamic and uncertain conditions. The model’s performance is validated using a real-world case study from the dairy industry, demonstrating its effectiveness in achieving sustainable supply chain objectives under varying demand scenarios and operational constraints. Comparative analyses with metaheuristic methods like NSGA-II further highlight the robustness and efficiency of the proposed approach.
Keywords:
Sustainable supply chain, Perishable dairy products, Demand prediction, multi-objective optimization, CNN-based routing.