A Hybrid Metaheuristic Approach for Multi-Objective Supply Chain Network Design under Uncertainty

This paper presents a hybrid metaheuristic framework for designing supply chain networks under uncertainty, optimizing multiple conflicting objectives simultaneously. The objectives considered are minimizing total cost, minimizing delivery time, and maximizing the robustness (resilience) of the network. Uncertainty in demand, transportation times, and facility disruptions is modelled via scenario‐based stochastic programming and robust optimization. The proposed hybrid method combines a Non-Dominated Sorting Genetic Algorithm II (NSGA-II) with Tabu Search (TS) for local refinement, enabling efficient exploration of the solution space. Computational experiments on publicly available benchmark instances and a realistic case study demonstrate that the hybrid method outperforms standard NSGA-II, NSGA-III, and Particle Swarm Optimization (PSO) in terms of Pareto frontier quality (hypervolume and spacing) and computational time. Results indicate that integrating local search (Tabu Search) improves robustness by up to ۱۵% while only increasing cost by ۳–۵%, under typical demand uncertainty. The proposed approach provides decision‐makers with a set of efficient trade‐off network designs, enabling more resilient supply chain configurations under uncertainty.This paper presents a hybrid metaheuristic framework for designing supply chain networks under uncertainty, optimizing multiple conflicting objectives simultaneously. The objectives considered are minimizing total cost, minimizing delivery time, and maximizing the robustness (resilience) of the network. Uncertainty in demand, transportation times, and facility disruptions is modelled via scenario‐based stochastic programming and robust optimization. The proposed hybrid method combines a Non-Dominated Sorting Genetic Algorithm II (NSGA-II) with Tabu Search (TS) for local refinement, enabling efficient exploration of the solution space. Computational experiments on publicly available benchmark instances and a realistic case study demonstrate that the hybrid method outperforms standard NSGA-II, NSGA-III, and Particle Swarm Optimization (PSO) in terms of Pareto frontier quality (hypervolume and spacing) and computational time. Results indicate that integrating local search (Tabu Search) improves robustness by up to ۱۵% while only increasing cost by ۳–۵%, under typical demand uncertainty. The proposed approach provides decision‐makers with a set of efficient trade‐off network designs, enabling more resilient supply chain configurations under uncertainty.

Keywords:

Hybrid Metaheuristic , Multi‐Objective Optimization , Supply Chain Network Design , Uncertainty , NSGA‐II Tabu Search , Robustness , Resilience