考虑订单可拆分的第三方共享制造平台产能匹配策略

doi:10.16381/j.cnki.issn1003-207x.2022.2182

摘要/Abstract

摘要：

为帮助制造企业缓解产能不足和产能过剩的困境，支持企业间产能共享的第三方制造平台相继出现并迅速发展。本文针对订单可拆分情形下第三方共享制造平台上的产能匹配问题，考虑产能供需数量、价格及最低匹配量等约束，以平台收益最大化为目标构建了产能拆分匹配模型，并综合贪婪算法（greedy algorithm，GA）和大邻域搜索算法（large neighborhood search algorithm， LNS）设计了求解模型的两阶段算法（GA-LNS）。结合现实数据构建多组数值算例进行分析，结果表明：（1）随着产能供方能接受的最低匹配量的降低，平台收益和产能供方总收益均正向增加；（2）与GUROBI及变邻域搜索算法求解结果相比，本文设计的GA-LNS算法具有更好的有效性；（3）GA-LNS求解的结果相对偏差指数（RDI）变化范围很小，GA-LNS算法具有较好的稳定性。本文所建立的模型和设计的算法能够在较短的时间内给出质量较高的解，帮助第三方共享制造平台更有效地匹配供需双方产能及提升制造资源利用率。

关键词: 第三方平台, 共享制造, 产能匹配, 订单拆分

Abstract:

With the development of the new generation of information technologies and shared economics， third-party shared manufacturing platforms （TPSMPs） are booming. Such platforms support the sharing of production capacity between manufacturing enterprises by integrating the dispersed and idle capacity of manufacturing enterprises in the market and matching capacity demand and supply. That is， manufacturing enterprises with overcapacity （capacity suppliers） in the TPSMP can share their surplus capacity， and manufacturing enterprises with insufficient production capacity （capacity demanders） can find suitable capacity suppliers and use their surplus capacity. In practice， considering the mismatch between capacity supply and demand in quantity， the order from one capacity demander on the TPSMP cannot usually be fully satisfied by a single capacity supplier， thus the capacity demander will allow the TPSMP to split its order to multiple capacity suppliers. Order splitting will incur additional cooperative costs， different capacity matching decisions will also affect the profitability of the TPSMP. Meanwhile， to guarantee that orders are fully satisfied， the TPSMP needs to decide not only which capacity suppliers to match with the capacity demanders， but also the matching quantities of the corresponding capacity suppliers， which complicates the TPSMP's capacity matching decisions.The main work of our paper includes the following three parts. First， focusing on the capacity matching problem of the TPSMP based on order splitting， the capacity split matching model maximizing the profit of the TPSMP is constructed considering the capacity of supply and demand quantity， price， minimum matching quantity， etc. Second， to solve the model， a two-stage heuristic method is designed combining greedy algorithm （GA） and large neighborhood search algorithm （LNS）. In the first stage， GA is applied to generate an initial feasible solution， in the second stage， LNS is adopted to improve the initial feasible solution. Finally， through numerical examples， the effectiveness of GA-LNS is verified， and some insights are obtained.The main results show that　（1） the revenue of the platform and the total revenue of the suppliers increase positively with the decrease of the minimum matching quantity acceptable for capacity suppliers；（2） the effectiveness of the two stage heuristic algorithm GA-LNS is verified by comparing with the results obtained by GUROBI and variable neighborhood search algorithm （VNS）；（3） the variation range of the relative deviation index （RDI） of the results obtained by GA-LNS is very small， and the GA-LNS algorithm has good stability. Our algorithm can obtain high quality solutions in a short time， which provides theoretical reference and decision-making basis for third-party shared manufacturing platform to solve the capacity matching problem considering order splitting.

Key words: third-party platform, shared manufacturing, capacity matching, order splitting

中图分类号:

C931

曹端阳,张旭梅,但斌. 考虑订单可拆分的第三方共享制造平台产能匹配策略[J]. 中国管理科学, 2025, 33(10): 225-235.

Duanyang Cao,Xumei Zhang,Bin Dan. Research on Capacity Matching Strategy of Third-party Shared Manufacturing Platform Considering Order Splitting[J]. Chinese Journal of Management Science, 2025, 33(10): 225-235.

图/表 8

表1

图1

表2

不同算例的GA-LNS算法优化效果"

算例	总收益(元)			产能利用率(%)
算例	$F 0$	$F G A - L N S$	$η$ (%)	$C U 0$	$C U G A - L N S$	$θ$
1	21142	29727	40.60	66.72	96.25	44.25
2	323214	387702	19.95	76.57	89.33	16.66
3	291740	369382	26.61	69.84	88.67	26.97
4	668252	728190	8.97	81.65	89.06	9.08
5	59034	61595	4.34	69.21	72.52	4.79
6	87703	99629	13.60	71.26	81.81	14.80
7	78768	82581	4.84	89.71	94.25	5.06
8	13608534	14096731	3.59	67.06	69.77	4.04
9	975870	1092961	12.00	86.17	96.48	11.97
10	1472017	1620734	10.10	82.50	91.17	10.51
11	99975	110175	10.20	79.48	87.25	9.77
12	140047	147385	5.24	72.06	76.09	5.59
13	5704075	6263378	9.81	86.95	95.54	9.88
14	92679	99946	7.84	88.63	95.75	8.04
15	4295845	4544538	5.79	87.37	92.12	5.43
16	1202491	1326336	10.30	80.19	89.06	11.06
17	11167970	11746442	5.18	88.61	93.29	5.29
18	45103390	46533241	3.17	84.06	86.10	2.42
19	2888583	3043971	5.38	81.29	85.11	4.70
20	24106717	25309585	4.99	87.05	91.42	5.01
平均	—	—	10.63	—	—	10.77

表2

表3

GA-LNS和GUROBI在不同测试实例的结果比较"

算例	目标函数(元)					计算时间(秒)
算例	$F ¯ *$	$F ¯ U B$	$F ¯ G A - L N S$	$η ¯ *$ (%)	$η ¯ U B$ (%)	$T ¯ G u r o b i$	$T ¯ G A - L N S$	$η ¯$ (%)
1	26887	26887	26887	0.00	0.00	5.71	0.28	-94.4
2	407731	407739	407719	0.00	0.00	14.46	0.35	-95.5
3	402822	402832	402811	0.00	-0.01	41.66	0.78	-98.0
4	583883	583891	583879	-0.00	-0.00	68.55	1.21	-95.7
5	77828	78283	77436	-0.49	-1.05	3419.01	3.93	-99.9
6	101423	101675	101194	-0.23	-0.48	3405.84	3.86	-99.9
7	80366	81058	80089	-0.35	-1.20	3600	5.29	-99.9
8	15611369	15614885	15576124	-0.23	-0.25	3600	10.85	-99.7
9	1111048	1119735	1096660	-1.29	-2.07	3600	10.90	-99.2
10	1598010	1609271	1585693	-0.78	-1.48	3600	17.93	-99.5
平均	—	—	—	-0.34	-0.65	—	—	-98.2

表3

表4

GA-LNS和VNS算法在不同测试实例下的结果比较"

算例	目标函数(元)			计算时间(秒)
算例	$F ¯ G A - L N S$	$F ¯ V N S$	$ζ ¯ 1$ (%)	$T ¯ G A - L N S$	$T ¯ V N S$	$ζ ¯ 2$ (%)
11	103828	100311	3.6	20.71	220.78	-90.1
12	153779	148745	3.4	28.93	170.11	-82.9
13	6175227	5864517	5.3	34.25	144.12	-75.5
14	99830	97077	2.9	22.40	226.38	-89.7
15	4482789	4254766	5.4	39.90	648.61	-93.7
16	1326678	1208674	9.9	165.73	622.91	-70.6
17	11706187	11317925	3.5	267.04	5125.92	-94.2
18	45556541	44102298	3.3	151.69	5064.80	-97.0
19	3097965	2927046	5.9	394.49	4161.21	-88.8
20	25457873	24825917	2.6	373.99	9061.04	-95.8
平均	—	—	4.58	—	—	-87.83

表4

图2

图3

表6

最低匹配量对匹配结果的影响"

算例	$q$	平台收益	产能供方总收益
1	500	29727	345311
	600	29696	344960
	700	29609	343949
	800	24586	283887
2	900	388305	4472403
	1000	387762	4466165
	1100	369542	4255484
	1200	365801	4212464
	1300	364258	4194718
	1400	364019	4191968
	1500	339487	3908701

表6

参考文献 35

[1]	国家信息中心分享经济研究中心课题组. 共享经济：从起步期向成长期加速转型——中国共享经济发展年度报告（2018）［J］. 新华月报， 2018（12）： 44-47.
	National Information Center Sharing Economy Research Center Research Group. Sharing economy： Accelerating the transformation from the initial stage to the growth stage：Annual report on the development of sharing economy in China （2018）［J］. Xinhua Monthly， 2018（12）： 44-47.
[2]	Yu C， Jiang X， Yu S， et al. Blockchain-based shared manufacturing in support of cyber physical systems： Concept， framework， and operation［J］.Robotics and Computer-Integrated Manufacturing，2020，64： 101931.
[3]	Yu C， Xu X， Yu S， et al. Shared manufacturing in the sharing economy： Concept， definition and service operations［J］. Computers & Industrial Engineering， 2020， 146： 106602.
[4]	余菲菲，高霞. 互联背景下我国制造企业生态化转型路径的选择机制研究：基于fsQCA方法［J］. 管理工程学报， 2020， 34（5）： 32-41.
	Yu F F， Gao X. Research on the path of ecological transformation of Chinese manufacturing enterprises under the background of iInternet： Based on the fsQCA method［J］. Journal of Industrial Engineering and Engineering Management， 2020， 34（5）： 32-41.
[5]	晏鹏宇，杨柳，车阿大. 共享制造平台供需匹配与调度研究综述［J］. 系统工程理论与实践， 2022， 42（3）： 811-832.
	Yan P Y， Yang L， Che A D. Review of supply-demand matching and scheduling in shared manufacturing［J］. Systems Engineering-Theory & Practice， 2022， 42（3）： 811-832.
[6]	Xu X. From cloud computing to cloud manufacturing［J］. Robotics and Computer-Integrated Manufacturing， 2012， 28（1）： 75-86.
[7]	Wu D， Greer M J， Rosen D W， et al. Cloud manufacturing： Strategic vision and state-of-the-art［J］. Journal of Manufacturing Systems， 2013， 32（4）： 564-579.
[8]	Ren L， Zhang L， Wang L， et al. Cloud manufacturing： Key characteristics and applications［J］. International Journal of Computer Integrated Manufacturing， 2017， 30（6）： 501-515.
[9]	李伯虎，张霖，任磊，等. 云制造典型特征、关键技术与应用［J］. 计算机集成制造系统， 2012， 18（7）： 1345-1356.
	Li B H， Zhang L， Ren L， et al. Typical characteristics， technologies and applications of cloud manufacturing［J］. Computer Integrated Manufacturing Systems， 2012， 18（7）： 1345-1356.
[10]	Zhang L， Luo Y， Tao F， et al. Cloud manufacturing： A new manufacturing paradigm［J］. Enterprise Information Systems， 2014， 8（2）： 167-187.
[11]	李伯虎，张霖，任磊，等. 再论云制造［J］. 计算机集成制造系统， 2011， 17（3）： 449-457.
	Li B H， Zhang L， Ren L， et al. Further discussion on cloud manufacturing［J］. Computer Integrated Manufacturing Systems， 2011， 17（3）： 449-457.
[12]	Huang B， Li C， Yin C， et al. Cloud manufacturing service platform for small- and medium-sized enterprises［J］. The International Journal of Advanced Manufacturing Technology， 2013， 65（9）： 1261-1272.
[13]	Wang X V， Xu X W. An interoperable solution for cloud manufacturing［J］. Robotics and Computer-Integrated Manufacturing， 2013， 29（4）： 232-247.
[14]	Ren L， Zhang L， Tao F， et al. Cloud manufacturing： From concept to practice［J］. Enterprise Information Systems， 2015， 9（2）： 186-209.
[15]	眭蓉华，张旭梅，但斌，等. 考虑交叉网络效应的第三方平台双边用户引入与增值服务策略［J］. 中国管理科学， 2023， 31（11）： 174-183.
	Sui R H， Zhang X M， Dan B， et al. Bilateral user introduction and value-added service strategies for the third-party platform with cross-network effects［J］. Chinese Journal of Management Science， 2023， 31（11）： 174-183.
[16]	Pan X Y， Ma J Z， Zhao D Z. Study on pricing behaviour and capacity allocation of cloud manufacturing service platform［J］. Cluster Computing， 2019， 22（6）： 14701-14707..
[17]	赵道致，陈慢慢. 考虑用户时间敏感的云制造平台定价策略研究［J］. 管理学报， 2021， 18（2）： 262-269.
	Zhao D Z， Chen M M. Research on pricing strategy of cloud manufacturing platform considering time-sensitive customers［J］. Chinese Journal of Management， 2021， 18（2）： 262-269.
[18]	Zhang X， Sui R， Dan B， et al. Bilateral value-added services and pricing strategies of the third-party platform considering the cross-network externality［J］. Computers & Industrial Engineering， 2021， 155： 107196.
[19]	Basu A， Bhaskaran S， Mukherjee R. An analysis of search and authentication strategies for online matching platforms［J］. Management Science， 2019， 65（5）： 2412-2431.
[20]	任磊，任明仑. 基于学习与协同效应的云制造任务动态双边匹模型［J］.中国管理科学，2018，26（7）： 63-70.
	Ren L， Ren M L. Dynamic two-sided matching method of cloud manufacturing task based on learning and synergy effect［J］. Chinese Journal of Management Science， 2018， 26（7）： 63-70.
[21]	Li B， Yang Y， Su J， et al. Two-sided matching model for complex product manufacturing tasks based on dual hesitant fuzzy preference information［J］. Knowledge-Based Systems， 2019， 186： 104989.
[22]	He L， Qian Z. Intent-based resource matching strategy in cloud［J］. Information Sciences， 2020， 538： 1-18.
[23]	Delaram J， Houshamand M， Ashtiani F， et al. A utility-based matching mechanism for stable and optimal resource allocation in cloud manufacturing platforms using deferred acceptance algorithm［J］. Journal of Manufacturing Systems， 2021， 60： 569-584.
[24]	Pu D， Yuan G. Two-sided matching model considering multi-information fusion of stakeholders［J］. Expert Systems with Applications， 2023， 212： 118784.
[25]	Tong H， Zhu J. New peer effect-based approach for service matching in Cloud Manufacturing under uncertain preferences［J］. Applied Soft Computing， 2020， 94： 106444.
[26]	Tong H， Zhu J. An RNN-MCDA method to support efficient and stable matching under uncertain preferences in large-scale sharing platforms［J］. Applied Soft Computing， 2022， 124： 108927.
[27]	Guo S， Dong M. Order matching mechanism of the production intermediation Internet platform between retailers and manufacturers［J］. The International Journal of Advanced Manufacturing Technology， 2021， 115（3）： 949-962..
[28]	Lim M K， Xiong W， Wang Y. A three-tier programming model for service composition and optimal selection in cloud manufacturing［J］. Computers & Industrial Engineering， 2022， 167： 108006.
[29]	Dell’Amico M， Delorme M， Iori M， et al. Mathematical models and decomposition methods for the multiple knapsack problem［J］. European Journal of Operational Research， 2019， 274（3）： 886-899.
[30]	王晶，刘昊天，黄钧. 考虑伤情分类的灾后创伤伤员救治与转运路径优化研究［J］. 中国管理科学， 2017， 25（8）： 114-122.
	Wang J， Liu H T， Huang J. Research on the route optimization of ambulance treatment and transportation after disaster based on the injured classification［J］. Chinese Journal of Management Science， 2017， 25（8）： 114-122.
[31]	Wang Y， Liu H， Peng B， et al. A three-phase matheuristic algorithm for the multi-day task assignment problem［J］. Computers & Operations Research， 2023， 159： 106313.
[32]	Zipfel B， Neufeld J， Buscher U. An iterated local search for customer order scheduling in additive manufacturing［J］.International Journal of Production Research， 2024， 62（3）： 605-625.
[33]	Brimberg J， Mladenović N， Todosijević R， et al. A basic variable neighborhood search heuristic for the uncapacitated multiple allocation p-hub center problem［J］. Optimization Letters， 2017， 11（2）： 313-327.
[34]	Ahn G， Park Y J， Hur S. Performance computation methods for composition of tasks with multiple patterns in cloud manufacturing［J］. International Journal of Production Research， 2019， 57（2）： 517-530.
[35]	李珍萍，贾顺顺，卜晓奇，等. 无人仓系统储位分配问题的优化模型与算法［J］. 中国管理科学， 2022， 30（1）： 124-135.
	Li Z P， Jia S S， Bu X Q， et al. The optimization model and algorithm for storage location assignment problem in unmanned warehouse system［J］. Chinese Journal of Management Science， 2022， 30（1）： 124-135.

文献	匹配类型			数量拆分	方法
文献	一对一	一对多	多对多	数量拆分	方法
任磊和任明仑^[20]	√				考虑学习与协同效应的双向匹配方法
Li等^[21]	√				基于投影技术的双边匹配方法
He和Qian^[22]	√				基于模糊聚类的双边匹配算法
Delaram等^[23]	√				延迟接受算法
Pu和Yuan^[24]	√				TODIM方法
Tong和Zhu^[25]		√			改进布谷鸟算法
Tong和Zhu^[26]		√			RNN-MCDA方法、改进的布谷鸟算法
Guo和Dong^[27]			√		多目标粒子群算法
Lim等^[28]			√		改进NSGA-II算法
本文			√	√	贪婪算法、大邻域搜索算法