B2C电子商务二级仓储网络订单履行的在线动态库存分配优化研究

doi:10.16381/j.cnki.issn1003-207x.2021.1215

摘要/Abstract

摘要：

为了保证B2C订单履行的就近发货，B2C的中心仓需要每天根据当天库存状态和需求信息对分仓进行动态补货。本文将这种补货称为在线动态库存分配。这种库存分配对减少转载运输和销售损失有重要影响，是B2C订单快速响应的重要支撑。本文为这类新的库存分配问题构建了新的模型和算法。首先，通过深入分析每天库存分配导致的当天和第二天的五种不同的缺货成本，构建了在线动态库存分配的单阶段决策模型。然后，利用边际缺货成本分析，推导出了不考虑运输约束时应优先考虑中心仓当天的需求，并为当天和第二天的需求分别设计了不同的均衡库存分配策略。接着，在考虑运输约束时，提出了基于平均缺货机会成本的迭代求解算法。最后，通过数值实验验证了模型和算法有效性。本文所提出的模型和算法拓展了库存分配的相关理论，为B2C订单履行的库存分配优化提供重要参考。

关键词: B2C订单履行, 二级仓储网络, 在线动态库存分配, 均衡分配策略, 迭代算法

Abstract:

To make the B2C order always be fulfilled by the closest warehouse， the Regional Distribution Center （RDC） of B2C have to replenish the inventory to Front Distribution Center （FDC） according to the available inventory and the daily demand. This process of the replenishment is called online dynamic inventory allocation， which could greatly reduce the inventory shortage and reduce the transshipment cost and the sales loss. So， the online dynamic inventory allocation become an important issue for the fastest response of the e-fulfillment. In this paper， the daily decision model and the solving algorithm of the online dynamic inventory allocation problem are presented. First， the cost of inventory shortage caused by the daily inventory allocation decision are analyzed， which includes five different shortage cost. And， the daily decision model is proposed based on these shortage costs. Then， according to the marginal analysis of inventory shortage cost， it is concluded that the RDC daily demand of the decision day should be satisfied preferentially. After that， two equilibrium inventory allocation strategies are presented specifically for the current day and the next day when transportation is unlimited. Furthermore， when both the limitation of quantity and the product item of transportation are considered， an iterative algorithm based on average opportunity cost of allocation is constructed. Finally， numerical experiments are implemented to simulate the inventory allocation operation of B2C e-commerce. The results of simulation demonstrate that both the model and the algorithm could obtain a sequence of optimal daily allocations in an acceptable computation time. The daily decision model and the solving algorithm of this study can help the B2C retailor develops better inventory allocation policies.

Key words: B2C order fulfillment, two-echelon warehouse network, online dynamic inventory allocation, equilibrium allocation strategy, iteration algorithm

中图分类号:

F253.4

熊浩,鄢慧丽. B2C电子商务二级仓储网络订单履行的在线动态库存分配优化研究[J]. 中国管理科学, 2024, 32(7): 248-257.

Hao Xiong,Huili Yan. The Optimization of Online Dynamic Inventory Allocation for Order Fulfillment in Two Echelon Warehouse Network of B2C e-Commerce[J]. Chinese Journal of Management Science, 2024, 32(7): 248-257.

图/表 12

图1

表1

模型符号定义"

变量		变量描述
决策变量	$x i t R$	商品i在RDC仓库第t天决策时的库存分配量（预留库存量）
	$x i j t$	商品i从RDC向FDC仓库j第t天决策时的库存分配量
	$y i j t$	商品i从RDC向FDC仓库j第t天决策时的库存分配量是否为0
	$w i j t$	商品i的在各个仓库的第t天决策时第二天的库存分配量
	$p i$	商品i的销售损失单位成本
	$q i$	商品i的紧急转载单位成本，且 $p i ≥ q i$
	$r i t$	商品i在RDC仓库第t天接收到来自供应商的补货量（第t天夜间抵达）
	$α j$	FDC仓库j顾客因本地缺货而放弃购买占未满足需求的比例
	N_j	FDC j线路每天可调拨商品种类个数上限
	M_j	FDC j线路每天可调拨商品总量上限
	$d i t R$	商品i在RDC第t天的需求（决策当天已知）
	$d ? i t + 1 R$	商品i在RDC第t+1天的需求（决策当天未知）
	$d i j t$	商品i在FDC j第t天的需求（决策当天已知）
	$d ? i j t$	商品i在FDC j第t+1天的需求（决策当天未知）
	$I i t R$	商品i在 RDC第t天的初始库存
	$I i j t$	商品i在 FDC j第t天的初始库存水平
	$b i j t$	第t天FDC j分配后缺货补货未满足的销售损失成
	$C i t R$	第t天RDC分配后的销售损失成本
	$C i j t F$	第t天FDC j期初的销售损失成本（受前t-1天的决策影响）
	$C i t Z$	第t天库存分配后，满足缺货补货的转载成本
	$C i t R F$	第t天库存分配后，缺货补货未满足的销售损失成本
	$C ? i j t + 1 R$	第t天决策时预估的第t+1天RDC的缺货销售损失成本
	$C ? i j t + 1 F$	第t天决策时预估的第t+1天FDC的缺货销售损失成本
	$z i t$	第t天库存分配后，商品i在各个FDC转载总量
	$v i j t$	决策当天的完成当天的库存分配之后的库存水平

表1

图2

图3

图4

表2

图5

图6

图7

图8

图9

表3

算法的数值实验结果对比"

序号	平均缺货成本	时间 (秒)	安全库存系数 k
1	4014.32	641.72	k1=k2=k3=k4=3;
2	4003.70	630.72	k1=k3=3.5, k2=k4=3
3	3998.31	642.42	k1=k2 =k3=k4=3.5
4	4013.48	646.62	k1=k3=4, k2=k4=3.5
5	4004.07	628.96	k1=k3=3.7, k2=k4=3.5
6	3996.47	636.82	k1=k3=3.3, k2=k4=3.5
7	3998.60	653.08	k1=k2=k3=k4=3.3
8	3996.67	636.60	k1=k3=3.3, k2=k4=3.4
9	3997.01	651.38	k1=3.2,k2=3.4; k3=3.3, k4=3.5
10	3995.42	646.42	k1=3.3,k2=3.5; k3=3.4, k4=3.6
11	3996.86	633.73	k1=3.3,k2=3.5; k3=3.5, k4=3.7
12	3996.28	631.92	k1=3.3,k2=3.5; k3=3.4, k4=3.7
13	3995.49	633.54	k1=3.3,k2=3.5;k3=3.5, k4=3.6
14	3996.38	642.04	k1=3.3,k2=3.4;k3=3.4,k4=3.6
15	3995.02	646.62	k1=3.2,k2=3.5; k3=3.4, k4=3.6
16	3996.01	638.98	k1=3.1, k2=3.5; FDC: k3=3.4, k4=3.6
17	3986.16	678.05	$k i j = G ˉ 0 - 1 θ α j p i$ ，对θ进行搜索

表3

参考文献 27

1	Clark A J， Scarf H. Optimal policies for a multi-echelon inventory problem［J］. Management science， 1960， 6（4）： 475-490.
2	De Kok A G. Hierarchical production planning for consumer goods［J］. European Journal of Operational Research， 1990， 45（1）： 55-69.
3	Lagodimos A G. Multi-echelon service models for inventory systems under different rationing policies［J］. International Journal of Production Research， 1992， 30（4）： 939-956.
4	Lagodimos A G， Koukoumialos S. Service performance of two-echelon supply chains under linear rationing［J］. International Journal of Production Economics， 2008， 112（2）： 869-884.
5	Van der Heijden M C， Diks E B， De Kok A G. Stock allocation in general multi-echelon distribution systems with （R， S） order-up-to-policies［J］. International Journal of Production Economics， 1997， 49（2）： 157-174.
6	Huang B， Iravani S M R. Optimal production and rationing decisions in supply chains with information sharing［J］. Operations Research Letters， 2007， 35（5）： 669-676.
7	Paul B， Rajendran C. Optimal and heuristic base-stock levels and rationing policy for a divergent supply chain［J］. International Journal of Industrial and Systems Engineering， 2010， 5（4）： 460-484.
8	Samii A B. Impact of nested inventory allocation policies in a newsvendor setting［J］. International Journal of Production Economics， 2016，181（12）： 247-256.
9	Wanke P， Alvarenga H， Correa H， et al. Fuzzy inference systems and inventory allocation decisions： exploring the impact of priority rules on total costs and service levels［J］. Expert Systems with Applications， 2017， 85（12）： 182-193.
10	Jimenez G H S， Rodrigues T F O， Dantas M M， et al. A dynamic inventory rationing policy for business-to-consumer e-tail stores in a supply disruption context［J］. Computers & Industrial Engineering， 2020，142：106379.
11	Wang Y， Zhang S H. Optimal production and inventory rationing policies with selective-information sharing and two demand classes［J］. European Journal of Operational Research， 2021，288（2）：394-407.
12	Protopappa-Sieke M， Sieke M A， Thonemann U W. Optimal two-period inventory allocation under multiple service level contracts［J］. European Journal of Operational Research， 2016， 252（1）： 145-155.
13	Edirisinghe C， Derek A. Lower bounding inventory allocations for risk pooling in two-echelon supply chains［J］. International Journal of Production Economics， 2017， 187（5）： 159-167.
14	Doğru M K， De Kok A G， van Houtum G J. A numerical study on the effect of the balance assumption in one-warehouse multi-retailer inventory systems［J］. Flexible Services and Manufacturing Journal， 2009， 21（3）： 114-147.
15	Tiemessen H G H， Fleischmann M， van Houtum G J， et al. Dynamic demand fulfillment in spare parts networks with multiple customer classes［J］. European Journal of Operational Research， 2013， 228（2）： 367-380.
16	Jackson P L， Muckstadt J A， Li Y X. Multiperiod stock allocation via robust optimization［J］. Management Science， 2018， 65（2）： 794-818.
17	Das C. Supply and redistribution rules for two-location inventory systems： One-period analysis［J］. Management Science， 1975， 21（7）： 765-776.
18	Jönsson H， Silver E A. Analysis of a two-echelon inventory control system with complete redistribution［J］. Management Science， 1987， 33（2）： 215-227.
19	Banerjee A， Burton J， Banerjee S. A simulation study of lateral shipments in single supplier， multiple buyers supply chain networks［J］. International Journal of Production Economics， 2003， 81（1）： 103-114.
20	Agrawal V， Chao X L， Seshadri S. Dynamic balancing of inventory in supply chains［J］. European Journal of Operational Research， 2004， 159（2）： 296-317.
21	Lee Y H， Jung J W， Jeon Y S. An effective lateral transshipment policy to improve service level in the supply chain［J］. International Journal of Production Economics， 2007， 106（1）： 115-126.
22	熊浩. 二级库存系统中库存策略的比较研究［J］. 系统工程理论与实践， 2007（6）： 100-104.
	Xiong H. Introduction of the relationship between safety stock and transshipment in two-echelon systems［J］. Systems Engineering-Theory & Practice， 2007（6）： 100-104.
23	Seidscher A， Minner S. A Semi-Markov decision problem for proactive and reactive transshipments between multiple warehouses［J］. European Journal of Operational Research， 2013， 230（1）： 42-52.
24	Feng P， Wu F， Fung R Y K， et al. Evaluation of two transshipment policies in a two-location decentralized inventory system under partial backordering［J］. Transportation Research Part E Logistics & Transportation Review， 2018， 118（10）： 207-224.
25	Van Wijk A C C， Adan I J B F， van Houtum G J. Optimal lateral transshipment policies for a two location inventory problem with multiple demand classes［J］. European Journal of Operational Research， 2019， 272（2）： 481-495.
26	Li Y， Liao Y， Hu X， et al. Lateral transshipment with partial request and random switching［J］. Omega， 2020，92：102134.
27	Zhang P， Xu X， Shi V， et al. Simultaneous inventory competition and transshipment between retailers［J］. International Journal of Production Economics， 2020，230：107781.

输入	数据名称
决策期间最初的初始库存	item_sku_id, dc_id, stock_quantity
外部供应商的补给计划	item_sku_id, date, replenish_quantity
需求的分布类型及参数	item_sku_id, dc_id, dist_type, para1, para2
商品的单位销售损失成本	item_sku_id, stockout_cost
FDC的单位商品的转载运输成本	q_j= 0.01 for all SKUs of all FDCs
FDC的运输品种约束	Nj=200 for all FDCs, j = 1,2,3,4,5
FDC的运输总量约束	M_j={3200,1600, 1200, 3600, 1600}, j = 1,2,3,4,5
缺货导致的销售损失率	α_j={1%,7%, 10%, 9%, 8%},j = 1,2,3,4,5