Investigating the impact of strategic warehouse design and product clustering on supply chain viability: A unified robust stochastic programming approach

Abstract

This study investigates the enhancement of supply chain (SC) viability through the integration of strategic warehouse design and product clustering under uncertainty. An integrated supply chain–warehouse design and inventory-distribution planning (ISWDIDP) problem is examined using a novel Unified Robust Stochastic Programming (URSP) model that leverages the strengths of both stochastic programming (SP) for known-unknown uncertainties and robust optimization (RO) for unknown-unknown uncertainties in customer demand. Solution strategies are developed using an Artificial Bee Colony Algorithm (ABCA) tailored to four distinct warehouse design strategies and two product clustering methods based on the K-means algorithm. A design of experiments (DoE) framework is employed to evaluate the impact of various controllable factors across case studies with different levels of demand variability. Multiple performance metrics—including overall cost, shortage cost, supplier and storage-area utilization cost, distribution cost, order receiving and picking cost, and storage-area utilization rate—are used to assess SC viability in terms of demand satisfaction, structural variety, process flexibility, and efficient redundancy. Moreover, a real-life case study based on a cardboard manufacturing factory is presented to validate the proposed approach in a practical setting. The findings underscore the critical role of strategic warehouse design and product clustering in enhancing SC viability under deep uncertainty, demonstrating that product clustering using both demand and product size features significantly improves performance compared to not clustering products.