Role of warehouse configuration in supply chain resilience under uncertainty: a stochastic programming approach with data-driven product clustering

This study investigates strategic warehouse configurations for enhancing supply chain resilience under demand and supply disruptions. A two-stage stochastic programming model is developed to jointly optimise warehouse configuration, inventory control, and distribution planning while minimising storage, transportation, picking, and shortage costs. Adopting a systems-oriented perspective, the proposed framework captures the interactions among suppliers, warehouse resources, inventory flows, and customer fulfilment activities within an integrated logistics system. Unlike deterministic approaches that rely on average demand and supply conditions, the proposed framework explicitly incorporates uncertainty in customer demand and supplier capacities. Due to the computational complexity of the problem, genetic algorithm-based solution approaches are employed to solve large-scale instances. Product clustering techniques, namely K-means and Gaussian Mixture Models (GMM), are employed to generate alternative storage assignment and warehouse configuration strategies. The effects of key controllable factors are evaluated using a structured design of experiments framework. A real-world case study demonstrates the effectiveness of the proposed approach, achieving cost reductions of up to 25% compared with current operations through lower shortage costs and improved order-handling efficiency. The results further indicate enhanced service levels, inventory performance, and supply chain resilience under volatile demand and supply conditions.