Logistics Optimization for Online Community Group Buying in Emerging O2O Business Modes
研究了在线社区团购中产品随机释放日期导致的物流效率问题，提出带随机释放日期和多产品的车辆路径模型，用分支定价算法求解，并通过美团案例验证了优化效果。
This paper addresses a critical logistics optimization challenge in the online community group buying (OCGB) business mode, where the stochastic release dates (SRDs) of products create inefficiencies in delivery planning. In general, vehicle routing models assume deterministic release dates (RDs), overlooking the uncertainty of RDs that is inherent in OCGB logistics. To address this shortcoming, we introduce a vehicle routing problem with SRDs and multiple products (VRP-SRD-M) aimed at minimizing total distance-related and penalty costs. The SRDs of aggregated products affects vehicle departure times, which poses computational challenges. We address this challenge by approximating SRDs with a Gumbel distribution and introducing a quality loss cost function to model overdue penalties. The problem is first formulated as an arc-flow model and then transformed into an equivalent set-partitioning model to increase computational efficiency and provide tighter upper bounds. To solve this problem, we propose a branch-and-price (B&P) algorithm based on the set-partitioning formulation, incorporating an efficient labelling algorithm to address the pricing problem (PP) and improve column generation (CG) strategies. Extensive computational experiments validate the advantages of incorporating SRDs in logistics optimization. Additionally, a real-world case study of Meituan's OCGB operations is used to quantify the impact of SRDs on distribution decisions, providing actionable managerial insights to increase delivery efficiency in stochastic environments.