Designing a centralised omnichannel distribution network

We investigate a centralised distribution network consisting of multiple retail stores and an online store. Each retail location faces unpredictable demand specific to its region, while the online store experiences varying demand across all locations. The warehouse primarily fulfils online orders, but retail stores can also use their inventory to support two omnichannel strategies: Ship-from-Store and Buy-Online-Pickup-in-Store. Before the selling season, retail stores must decide whether to implement one or both strategies, as they require prior preparation and resource allocation. Retail stores replenish their inventory from the distribution centre before the season starts. During the selling season, decision-makers optimise inventory allocation across the network to meet both in-store and online demand, aiming to maximise net profit. This problem is formulated as a two-stage stochastic optimisation model. We apply meta-heuristic techniques, including Simulated Annealing, Tabu Search, and Genetic Algorithm, to solve the model. Computational experiments demonstrate the effectiveness of these methods, achieving high-quality solutions with minimal deviation from the optimal outcome.