Abstract

With the rapid expansion of e-commerce platforms, improving inventory management efficiency and cost control for massive product portfolios has emerged as a critical operational challenge for merchants. This study addresses the multi-objective replenishment decision optimization problem for diverse product categories in e-commerce fulfillment warehouses by proposing a collaborative optimization model that integrates multi-objective integer programming with the CRITIC weighting method. The model aims to achieve equilibrium among three objectives: minimizing total operational costs, maximizing customer service levels, and minimizing inventory turnover days. A periodic inventory review strategy, dynamically linked to historical demand data, establishes an optimization framework with inventory lower and upper bounds as decision variables. To eliminate subjectivity in multi-objective weight allocation, the CRITIC weighting method is introduced to quantify the contrast intensity and conflict among objectives. Information entropy, calculated through standard deviation and Pearson correlation coefficients, enables objective weight assignments for cost, service level, and inventory efficiency targets. Simulation experiments utilizing operational data from JD Logistics (covering 37 merchants, 57 warehouses, and 2,303 products) demonstrate that the model dynamically optimizes replenishment quantities and inventory thresholds within 15-day cycles. Results reveal an 18.2% reduction in total operational costs, a 96.4% demand fulfillment rate, and an average inventory turnover time shortened to 1.3 days. This research provides theoretical and practical insights for intelligent decision-making in e-commerce supply chains. Future work may enhance the model’s dynamic responsiveness and global optimization capabilities through integration with intelligent optimization algorithms