Improved bisecting k-means algorithm-based order grouping method

Abstract

The invention discloses an improved bisecting k-means algorithm-based order grouping method. An improved bisecting k-means algorithm is adopted for solving the order batching problem of a distributioncenter; and the k-means clustering algorithm is improved from three aspects of k value selection, initial central value selection and abnormal point processing, so that the algorithm is prevented from falling into local optimum, a solved order grouping scheme is more effective and reasonable, and the order sorting efficiency is effectively improved.

Description

Technical field [0001] The invention relates to a grouping method, in particular to an order grouping method based on an improved dichotomous k-means algorithm. Background technique [0002] With the development of e-commerce, e-commerce will receive a large number of orders every day, and these orders are characterized by small batches, multiple varieties, and multiple batches. For these large-scale orders, e-commerce logistics centers are under increasing pressure. [0003] Order grouping is to group the collected customer orders according to specific rules, and arrange the orders of the same group on the same workbench for sorting, so as to shorten the order picking time and improve the picking efficiency. Currently, order grouping strategies are: 1. Seed algorithm, 2. Saving algorithm, 3. Priority rule algorithm, 4. Heuristic algorithm, 5. Data mining algorithm. [0004] The initial seed data of the seed algorithm is difficult to find a better value, and it is difficult to find...

AI Technical Summary

Problems solved by technology

[0004] It is difficult to find a better value for the initial seed data of the seed algorithm, and it is difficult to find the optimal grouping scheme for large-scale orders; the optimal rule algorithm is to classify customer orders and group them according to the priority of the order, but it does not consider the order The correlation between the obtained grouping schemes often cannot effectively reduce the sorting efficiency; and the general heuristic algorithm cannot be applied to large-scale order collections; for data mining algorithms, there are mainly two types of order grouping problems at this stage: Algorithm: association rule mining, k-means clustering algorithm

These two algorithms are suitable for large-scale order grouping problems, but the current order grouping scheme based on the k-means algorithm does not have three limitations of the reasonable improvement algorithm: 1. Determination of the k value, 2. Determination of the initial center, 3. The processing of abnormal data points makes the current scheme fail to effectively improve the sorting efficiency of the system

Images