关键词: C-均值算法, 分拣调度, 复合式分拣系统, 聚类, 蚁群算法

Abstract: In order to improve the whole sorting efficiency of automatic compositing sorting system in distribution center, a clustering method is proposed based on the similarity of stock keeping units (SKUs) and the balancing principle of sorting tasks. Compared with the traditional manual assignment of sorting scheduling policy, the method proposed not only solves the sorting-channel-frequent-exchange problem caused by the inconsistent of SKUs between different distribution routes, but also solves the imbalance of working tasks among multiple sorting systems. When the up-down hybrid picking system is used for the single distribution route, the tasks between the up-order-picking-line and down-order-picking-line, as well as the tasks of different channels within a single line, are often not balanced. So the picking task model is developed to minimize the difference of total picking amount between up-order-picking-line and down-order-picking-line and the mean square deviation among the picking tasks of different channels. In order to get the optimal assignment of sorting tasks, picking orders in the same distribution route are grouped by the distribution sequence and ant colony optimization (ACO) is applied to solve the model. Simulation result indicates that the sorting tasks between up-order-picking-line and down-order-picking-line can be assigned in terms of efficiency ratio which satisfies the working demands well in both efficiency and effectiveness.

Key words: Ant colony optimization, Automatic compositing sorting system, Clustering, C-means, Sorting scheduling policy