Research And Application Of Data-driven Automated Warehouse Picking Optimization Algorithm

With the rapid popularization of Internet technology and the rapid development of economic globalization,import cross-border e-commerce shows strong vitality,and user’s requirements for online shopping experience are gradually increasing.The traditional logistics warehouse employs manual sorting,which greatly increases the response time of the logistics warehouse.In order to better improve customers’ experience and save labor costs,automated picking operations have become a trend in the development of warehouse distribution centers.However,due to the scattered,large number and diversified types of orders,the automatic picking equipment cannot play its role well.Therefore,the research on the optimization algorithm of automatic storage picking is particularly important.This article investigates and analyzes the automated picking equipment introduced based on a typical cross-border e-commerce company,and aims to improve the picking efficiency of order through a reasonably designed automated warehouse picking optimization algorithm.The main research contents and results of this paper are as follows:(1)Research and algorithm design of goods placement.This paper analyses the characteristics of the storage bins of automated warehousing and picking equipment,traverses historical order data,and uses the FP-growth algorithm to dig out the relationship between the goods.Then,according to the volume of the bin and the loadbearing capacity of the robot vehicle,goods with a strong relationship are placed together,which can effectively reduce the I/O number of bins of automated picking equipment and save picking time.(2)Research and algorithm design of location allocation optimization.Through the analysis and research of automated picking equipment introduced by a typical crossborder e-commerce company,this paper mines the historical order data.At the same time,the sales volume of various commodities was counted,the turnover situation was analyzed and finally the location optimization goals and constraint conditions were built,and established the storage location model of automatic picking equipment.In this paper,the evolution reversal operator and parameter self-adaptive are introduced into the genetic algorithm to solve the objective of location optimization.Compared with the location allocation strategy based on the principle of turnover location correspondence,the feasibility of the algorithm is verified by the experimental data,and the suggestions of using different location optimization strategies in different scenarios are given,which provides ideas for the actual application of location optimization.(3)Research and algorithm design of order batch optimization problem.After the investigation and analysis of order picking operations of a typical cross-border ecommerce,an order batching model based on the information of commodity storage bins was established.According to the difference of order changes before and after the order was cut,a two-stage processing method was proposed and then an improved algorithm for different processing methods was put forward.Before the order is cut,orders will continue to flow in randomly and the number of orders is variable.This article proposes a new seed selection mode,which introduces the frequent purchase mode of goods,selects the order containing the most frequently purchased goods as the seed order and redefines the distance scale between orders and the seed order.After the order is cut,the number of orders will no longer change.This article proposes an improved K-Means order batching algorithm based on the idea of seed orders.This method redefines the cluster center and the distance from the order to the cluster center,uses the distance from the order to the cluster center as a clustering index,and solves the model.This study verified the company’s order data examples and compared the company’s current order processing methods and the first-come-first-served(FCFS)algorithm commonly used in traditional logistics.The experimental results show that the efficiency has increased by 13% and 15% respectively.This paper also compares the picking efficiency before the optimization of the location and order batching.The picking efficiency is increased by 33%,which proves the practicability of the algorithm.