A sustainable supply chain network considering lot sizing with quantity discounts under disruption risks: centralized...

Area of Science:

• Operations Research
• Supply Chain Management
• Sustainable Manufacturing

Background:

• Sustainable supply chain networks face challenges from lot-sizing, quantity discounts, and disruption risks.
• Integrating the triple bottom line (cost, environment, jobs) is crucial for holistic sustainability assessments.
• Transformer production presents a specific application context for supply chain optimization.

Purpose of the Study:

• To propose a novel framework for sustainable supply chain networks incorporating lot-sizing and disruption risks.
• To develop and compare centralized and decentralized optimization models for supplier selection and order allocation.
• To enhance inventory planning using an economic order quantity model for uncertainty management.

Main Methods:

• Formulation of a scenario-based stochastic mixed-integer non-linear programming approach for the centralized model.
• Development of a Stackelberg game model to analyze supplier competition and pricing under sustainability constraints.
• Application of a hybrid imperialist competitive algorithm (ICA) and particle swarm optimization (PSO) metaheuristic for solving the optimization models.

Main Results:

• The decentralized Stackelberg game model yields lower total costs compared to non-bi-level approaches, despite increased computational time.
• The hybrid ICA-PSO algorithm demonstrates robust performance, with PSO and ICA-PSO showing favorable standard deviations.
• Numerical results confirm the efficiency and performance of the proposed framework and the hybrid metaheuristic.

Conclusions:

• The proposed framework effectively addresses sustainability criteria and disruption risks in supply chain networks.
• The hybrid ICA-PSO algorithm provides a viable method for solving complex supply chain optimization problems.
• The study highlights the trade-off between solution optimality and computational efficiency in centralized versus decentralized models.