A Discrete Simulation-Based Optimization Algorithm for the Design of Highly Responsive Last-Mile Distribution Networks

Online and omnichannel retailers are proposing increasingly tight delivery deadlines, moving toward instant on-demand delivery. To operate last-mile distribution systems with such deadlines efficiently, defining the right strategic network design is of paramount importance. However, this problem exceeds complexity traditional networks for two main reasons: (1) reduced time available order handling (2) absence a cut-off that clearly separates collection periods. This renders state-of-the-art models inappropriate, as they assume periodic fulfillment based on cutoff. In study, we propose metamodel simulation-based optimization (SO) approach to strategically deadlines. Our methodology integrates an in-depth simulator techniques by extending black-box SO algorithm analytical model captures underlying structure decision problem. Based numerical study inspired efforts global fashion company introduce in Manhattan, show our outperforms contemporary approaches well deterministic stochastic programming methods. particular, method systematically yields designs superior expected cost performance. Furthermore, it converges good solutions lower computational budget more consistent finding high-quality solutions. We how congestion effects processing orders at facilities negatively impact performance through late potential consolidation. addition, sensitivity optimal increases become tight.