Multi-Depot Vehicle Routing Problem with Collaborative Replenishment Using ALNS–ABC Algorithm

In the context of multi-depot operations in the fast-moving consumer goods (FMCGs) industry, traditional logistics approaches increasingly fail to meet integrated requirements for cost, speed and environmental impact. Multi-depot collaboration, key to enhancing efficiency, faces challenges in route optimization, inventory management and carbon constraints. To balance economy and sustainability, this paper addresses multi-depot distribution challenges via the collaborative delivery vehicle routing problem with multi-depots and replenishment (CDVRP-MDR). It constructs a mixed-integer linear programming model with vehicle capacity and carbon constraints, then designs an adaptive large neighborhood search–artificial bee colony algorithm (ALNS–ABC). Through dynamic destruction-repair operator selection and gene reverse-order recombination, ALNS–ABC optimizes local search to balance global exploration and local development. Benchmark experiments confirm its effectiveness in multi-depot optimization. For FMCG firms’ centralized inventory issues (delayed transfers, high long-haul costs), it boosts inventory turnover and cross-depot response via collaborative transfers, offering a practical paradigm for green upgrading of multi-depot logistics systems.