基于不确定需求的货位分配策略

doi:10.16381/j.cnki.issn1003-207x.2022.0791

摘要/Abstract

摘要：

货位分配和货品拣选是仓库管理的两个重要方面，合理的货位分配策略能有效提高货品的拣选效率，节约仓储成本。常用的ABC货位分配策略将货物按品种和周转率进行分区分类。然而，在需求不确定情形下，考虑到安全库存的需要，单种货品在补货周期内有少部分货物的拣选为小概率事件，ABC货位分配策略会使得总周转频率高的这部分货物同样占据着距离I/O点较近货位，其周转频率低，从而导致了总拣货距离的增加。为此，本文提出了一种基于不确定需求的货位分配策略，通过考虑单种货品不同数量货物需求概率的不同，以需求概率进行分区，将同种货品进行拆分存放，使其分布在不同分区。各类货品的小概率需求部分会被存放在远离I/O点的货位，以此提高靠近I/O点货位的总需求概率，从而使存储和拣货过程更高效。首先，以平均单向拣货距离为优化目标，推导了考虑需求不确定的货品总周转率和距离表达式。其次，通过数值实验，将本文提出的策略同传统ABC货位分配策略进行拣货距离对比，实验结果表明，在不同需求曲线、标准差、服务水平下，基于需求概率（probability-of-retrieval，POR）的货位分配策略均在不同程度上优于传统ABC货位分配策略；在20%货品品类贡献80%需求的ABC斜率曲线下，在不同的需求概率分区策略里，第一分区（最靠近I/O点的区域）内货品累积需求概率在20%左右时，总拣货距离均相对最短。

关键词: 货位分配, 需求不确定, 分区拣选, 存储策略

Abstract:

In logistics supply chain management， the warehouse as the storage place of goods， its storage optimization has been the focus of research by warehouse managers and researchers. The allocation of storage positions and goods picking are two important aspects of inventory management. If the space allocation strategy is appropriate， the efficiency of goods picking can be effectively improved. The ABC allocation strategy commonly used in warehouse storage classifies goods by category and turnover. This allocation method does not consider the same kind of goods for split placement. However， when demand is uncertain， considering the need of safe stock， the selection of a few items in the replenishment cycle of a single goods is a low probability event. The ABC allocation strategy will make the items with a low probability of being selected in the goods with a high total turnover frequency also occupy the position closer to I/O point. The low turnover frequency leads to an increase in total picking distance and a decrease in picking efficiency.In this paper， a new allocation strategy based on uncertain demand is proposed. By considering the different demand probability of different quantity items in a single goods， the strategy divides the same goods into different areas. The low-probability demand portion of the goods will be stored away from the I/O point. In this way， the total demand probability near the I/O point is increased， thus making the storage and picking process more efficient. For the convenience of description and understanding， the goods location allocation strategy proposed in this paper is defined as the Probability-of-Retrieval（POR） goods location allocation strategy. The ABC location allocation strategy sorts goods according to turnover rate， but the POR strategy allocates the same goods according to the different demand probability under the different quantity in the replenishment period， and determines a probability boundary as the partition boundary， all kinds of goods are stored randomly within the same boundary.The average one-way picking distance is taken as the optimization goal. Firstly， the optimal order quantity of each goods is gotten by combining EOQ replenishment strategy. Secondly， according to the assumption of normal distribution and the POR location allocation strategy， the expressions of storage space， weighted turnover frequency and picking distance of goods in different subareas and the expressions of picking distance of ABC location allocation under uncertain demand are derived. Finally， Matlab 2018A is used to carry out numerical experiments， and the optimized results are compared under the different variable parameters such as demand curve， standard deviation， service level and so on， the influence of the change of the probability partition on the total picking distance is discussed.The experimental results show that under different parameters， the POR location allocation strategy is better than the traditional ABC location allocation strategy to varying degrees. The smoother the demand curve is， the smaller the demand difference of various goods is. The model based on demand probability proposed in this paper can save the total travel distance more than the traditional model， but at the same time， the increase of optimization degree is also more and more gentle； Under the POR strategy， too high service level is not appropriate； The smaller the standard deviation of goods demand distribution， the higher the degree of optimization； The more probability zones， the smaller the travel distance. Under the common ABC slope curve of 20% goods contributing 80% demand， in each probability partition strategy， when the cumulative demand probability of goods in the first partition （the area closest to the I/0 point） is about 20%， the total picking distance is relatively shortest.In the actual storage， the warehouse manager can estimate the demand distribution according to the previous orders， get the average value and standard deviation of different kinds of goods， and arrange the goods distribution better under the partition storage allocation strategy based on the demand probability.

Key words: allocation of storage positions, demand uncertainty, partition picking, storage strategy

中图分类号:

C934

邓旭东,黎婷,马云峰, 等. 基于不确定需求的货位分配策略[J]. 中国管理科学, 2026, 34(5): 166-174.

Xudong Deng,Ting Li,Yunfeng Ma, et al. Allocation Strategy Based on Uncertain Demand[J]. Chinese Journal of Management Science, 2026, 34(5): 166-174.

图/表 10

图1

图2

表1

主要参数符号"

参数符号	含义
$i ∈ 1,2, 3, …, N$	第 $i$ 项货品的索引号，数字越小，该类货品的需求频率越高
$k, k' ∈ 1,2, 3, …, n$	仓储中的分区号，数字越小，该分区越靠近I/O点，其中 $k$ 为POR策略下的分区号， $k'$ 为ABC分类法分区号
$N$	仓库内货品种类总数
$N k$	第 $k$ 分区内的货品种类数
$D i$	第 $i$ 项货品的年需求量
$Q i$	第 $i$ 项货品的经济订货批量
$P k i$	第 $i$ 项货品前 $k$ 个分区的货品分布量被访问到的累积需求概率
$P k$	前 $k$ 个分区中的货品分布量在总库区货品需求下被访问到的累积需求概率
$Λ (k)$	第 $k$ 区货品的总周转频率，即储存在 $k$ 区的所有货品在单位时间内的周转量
$Q i k$	第 $i$ 项货品在第 $k$ 分区的存储量
$t k$	储存/检索 $k$ 区单位货品的平均单程拣货距离
$T n$	分区为 $n$ 的仓库单位货品的平均单向总拣货距离

表1

图3

表2

不同需求曲线下两种策略的取货行程距离对比"

需求曲线形状因子 $s$	累积货品种类分区	[30%，50%，97%]
需求曲线形状因子 $s$	累积货品种类分区	$T p$	$T f$	优化程度 $I$ （%）
1（20%/20%)	100	105.62	277.11	61.89
0.748（20%/30%)	73,100	104.56	271.25	61.45
0.569(20%/40%)	26,94,100	101.75	248.13	58.99
0.431(20%/50%)	9,49,99,100	97.41	217.43	55.2
0.317(20%/60%)	4,30,81,100	91.34	180.66	49.44
0.222(20%/70%)	2,21,66,100	83.32	140.31	40.62
0.139(20%/80%)	1,12,50,86,100	72.35	97.47	25.77
0.065(20%/90%)	1,10,37,70,100	55.89	54.82	-1.95

表2

图4

图5

图6

图7

表3

不同概率分区对Tp的影响"

$k$ =3		$k$ =4		$k$ =5
累积概率分区	$T p$	累积概率分区	$T p$	累积概率分区	$T p$
[0.1,0.65,0.97]	76.88	[0.1,0.6,0.8,0.97]	58.19	[0.1,0.55,0.7,0.85,0.97]	48.29
[0.2,0.65,0.97]	71.79	[0.2,0.6,0.8,0.97]	54.65	[0.2,0.55,0.7,0.85,0.97]	46.25
[0.3,0.65,0.97]	73.70	[0.3,0.6,0.8,0.97]	57.93	[0.3,0.55,0.7,0.85,0.97]	50.86
[0.4,0.65,0.97]	80.83	[0.4,0.6,0.8,0.97]	66.37	[0.4,0.55,0.7,0.85,0.97]	60.57
[0.5,0.65,0.97]	92.85	[0.5,0.6,0.8,0.97]	79.69	[0.5,0.55,0.7,0.85,0.97]	75.15
[0.6,0.65,0.97]	110.09	[0.2,0.3,0.8,0.97]	64.02	[0.2,0.4,0.7,0.85,0.97]	44.61
[0.2,0.3,0.97]	108.61	[0.2,0.4,0.8,0.97]	55.94	[0.2,0.5,0.7,0.85,0.97]	44.46
[0.2,0.4,0.97]	91.88	[0.2,0.5,0.8,0.97]	52.79	[0.2,0.6,0.7,0.85,0.97]	49.38
[0.2,0.5,0.97]	80.11	[0.2,0.6,0.8,0.97]	54.62	[0.2,0.5,0.6,0.85,0.97]	46.83
[0.2,0.6,0.97]	73.24	[0.2,0.7,0.8,0.97]	62.28	[0.2,0.5,0.8,0.85,0.97]	48.72
[0.2,0.7,0.97]	71.87	[0.2,0.5,0.6,0.97]	65.59	[0.2,0.5,0.7,0.8,0.97]	46.50
[0.2,0.8,0.97]	77.73	[0.2,0.5,0.7,0.97]	58.08	[0.2,0.5,0.7,0.9,0.97]	45.17

表3

参考文献 25

[1]	李建斌，蒙铭友，戴宾. 电子商务环境下的存储策略优化研究：固定存储还是分类随机存储［J］. 中国管理科学， 2021， 29（8）： 67-80.
	Li J B， Meng M Y， Dai B. A study on storage policy optimization： Dedicated storage vs class-based random storage［J］. Chinese Journal of Management Science， 2021， 29（8）： 67-80.
[2]	陈璐，陆志强. 自动化立体仓库中的储位分配及存取路径优化［J］. 管理工程学报， 2012， 26（1）： 42-47.
	Chen L， Lu Z Q. Optimization for storage location assignments and interleaving problems in an automated storage/retrieval system［J］. Journal of Industrial Engineering and Engineering Management， 2012， 26（1）： 42-47.
[3]	王疏影，张益农，冯冰清，等. 基于AHP和运筹学思想的储位管理研究［J］. 制造业自动化， 2021， 43（9）： 101-105+121.
	Wang S Y， Zhang Y N， Feng B Q， et al. Research on storage management based on AHP and operations research［J］. Manufacturing Automation， 2021， 43（9）： 101-105+121.
[4]	Petersen C G， Schmenner R W. An evaluation of routing and volume-based storage policies in an order picking operation［J］. Decision Sciences， 1999， 30（2）： 481-501.
[5]	Eynan A， Rosenblatt M J. Establishing zones in single-command class-based rectangular as/rs［J］. IIE Transactions， 1994， 26（1）： 38-46.
[6]	宁浪，张宏斌，张斌. 面向JIT制造的零部件配送中心货位优化研究［J］.管理科学学报，2014，17（11）： 10-19.
	Ning L， Zhang H B， Zhang B. Optimizing the storage location assignment in a JIT-oriented parts distribution center［J］. Journal of Management Sciences in China， 2014， 17（11）： 10-19.
[7]	Hausman W H， Schwarz L B， Graves S C. Optimal storage assignment in automatic warehousing systems［J］. Management Science， 1976， 22（6）： 629-638.
[8]	Le-Duc T， De Koster R M B M. Travel distance estimation and storage zone optimization in a 2-block class-based storage strategy warehouse［J］. International Journal of Production Research， 2005， 43（17）： 3561-3581.
[9]	Petersen C G， Aase G R， Heiser D R. Improving order-picking performance through the implementation of class-based storage［J］. International Journal of Physical Distribution & Logistics Management， 2004， 34（7）： 534-544.
[10]	Rao S S， Adil G K. Optimal class boundaries， number of aisles， and pick list size for low-level order picking systems［J］. IIE Transactions， 2013， 45（12）： 1309-1321.
[11]	Guo X， Yu Y， De Koster R B M. Impact of required storage space on storage policy performance in a unit-load warehouse［J］. International Journal of Production Research， 2016， 54（8）： 2405-2418.
[12]	陈园园，吕志军，李宏亮，等. 基于随机存储策略的四向车密集仓储系统作业性能评估［J］. 东华大学学报（自然科学版）， 2021， 47（5）： 80-88.
	Chen Y Y， Lyu Z J， Li H L， et al. Operation performance evaluation of four-way shuttle high dense storage and retrieval system based on random storage strategy［J］. Journal of Donghua University （Natural Science）， 2021， 47（5）： 80-88.
[13]	郭晓龙. 仓储系统中存储策略与设备维护策略研究［D］. 合肥：中国科学技术大学， 2014.
	Guo X L. A study of storage policy and maintenance support strategy in warehousing systems［D］. Hefei： University of Science and Technology of China， 2014.
[14]	Goetschalckx M， Ratliff H D. Shared storage policies based on the duration stay of unit loads［J］. Management Science， 1990， 36（9）： 1120-1132.
[15]	Heskett J L. Cube- per- order index： A key to warehouse stock location ［J］. Transportation and Distribution Management， 1963（3）： 27-31.
[16]	Petersen C G， Aase G. A comparison of picking， storage， and routing policies in manual order picking［J］. International Journal of Production Economics， 2004， 92（1）： 11-19.
[17]	Ang M， Lim Y F. How to optimize storage classes in a unit-load warehouse［J］. European Journal of Operational Research， 2019， 278（1）： 186-201.
[18]	Heragu S S， Du L， Mantel R J， et al. Mathematical model for warehouse design and product allocation［J］. International Journal of Production Research， 2005， 43（2）： 327-338.
[19]	李珍萍，卜晓奇，陈星艺. 基于商品关联度的智能仓库储位分配问题研究［J］. 数学的实践与认识， 2020， 50（5）： 23-31.
	Li Z P， Bu X Q， Chen X Y. Research on storage allocation problem of intelligent warehouse based on correlation degree of items［J］. Mathematics in Practice and Theory， 2020， 50（5）： 23-31.
[20]	李明琨，蒋欣颖. 基于物料周转率与需求相关性的储位分配优化策略与算法［J］. 运筹与管理， 2018， 27（9）： 22-32.
	Li M K， Jiang X Y. A method for storage location assignment considering COI and demand correlations［J］. Operations Research and Management Science， 2018， 27（9）： 22-32.
[21]	蔺宇，邓永锋. 基于均衡化的单量份拣选储位分配方法［J］. 制造业自动化， 2013， 35（16）： 127-130.
	Lin Y， Deng Y F. The method of storage assignment in set parts supply based on the balance［J］. Manufacturing Automation， 2013， 35（16）： 127-130.
[22]	李诗珍. 基于工作量均衡的分区同步拣货系统储位分配与评价［J］. 包装工程， 2010， 31（11）： 114-118.
	Li S Z. Location assignment and evaluation in a synchronized zone picking system base on workload balance［J］. Packaging Engineering， 2010， 31（11）： 114-118.
[23]	胡少龙，胡志华，曹杨. 基于多候选储位的存取路径优化问题研究［J］. 运筹与管理， 2013， 22（5）： 111-116+165.
	Hu S L， Hu Z H， Cao Y. Optimization for storage or retrieval routing problem based on mutil-candidates storages location［J］. Operations Research and Management Science， 2013， 22（5）： 111-116+165.
[24]	李诗珍. 拣货方式、存储策略与路径策略协同研究［J］. 工业工程， 2011， 14（2）： 37-43.
	Li S Z. Synergistic effect study on picking， storing and routing policies［J］. Industrial Engineering Journal， 2011， 14（2）： 37-43.
[25]	Yu Y， de Koster R B M， Guo X. Class-based storage with a finite number of items： Using more classes is not always better［J］. Production and Operations Management， 2015， 24（8）： 1235-1247.