目的 为提高需求快速变化、波动较大的在线零售企业的仓库货位优化效率。方法 利用数据仓库和数据挖掘法，研究基于复合规则的动态货位指派策略。对该货位指派的数据集成分析、指标计算、规则生成和货位指派等4个模块进行分析，并设计库区标定算法和规则生成算法来生成货位指派规则集。结果 基于复合规则动态货位指派不仅能够节约拣货距离，而且拣货效率受需求变化的影响非常小。 结论 数值实验表明，与传统的货位指派策略相比，基于复合规则动态货位的指派系统可以得到更好的结果，并且在平均订单规模较大和需求偏度大的情况下效果更加明显。

The work aims to improve the optimization efficiency of storage location of the online retail business subject to the rapid change in demand and large fluctuation. The composite rules-based dynamic storage location assignment strategy was studied based on the data warehouse and data mining technology. The date integration analysis, index calculation, rule generation and storage location assignment of such storage location assignment were analyzed. Moreover, the zone calibration algorithm and rule generation algorithm were designed to generate the storage location assignment rule set. The composite rules-based dynamic storage location assignment could save the picking distance, and its picking efficiency was insensitive to the change in demand. Numerical experiments show that, compared with the traditional storage location assignment strategy, the composite rules-based dynamic storage location assignment system can get better results, and especially in the larger-sized average order with high demand skewness, the effect is more obvious.