Design and application of Internet of thingsbased warehouse management system for smart logistics 基于物联网的智能物流仓储管理系统的设计与应用
C.K.M. Lee, Yaqiong Lv, K.K.H. Ng, William Ho & K.L. Choy C.K.M. Lee、Yaqiong Lv、K.K.H. Ng、William Ho 和 K.L. Choy
To cite this article: C.K.M. Lee, Yaqiong Lv, K.K.H. Ng, William Ho & K.L. Choy (2018) 引用本文：C.K.M. Lee, Yaqiong Lv, K.K.H. Ng, William Ho & K.L. Choy (2018)
Design and application of Internet of things-based warehouse management system for smart logistics, International Journal of Production Research, 56:8, 2753-2768, DOI: 10.1080/00207543.2017.1394592 基于物联网的智能物流仓储管理系统的设计与应用，《国际生产研究杂志》，56：8，2753-2768，DOI：10.1080/00207543.2017.1394592
Design and application of Internet of things-based warehouse management system for smart logistics 基于物联网的智能物流仓储管理系统的设计与应用
C.K.M. Lee ^(a){ }^{\mathrm{a}}, Yaqiong Lv^(b**)\mathrm{Lv}^{\mathrm{b} *}, K.K.H. Ng^(a)\mathrm{Ng}^{\mathrm{a}}, William Ho^(c)\mathrm{Ho}^{\mathrm{c}} and K.L. Choy ^(a){ }^{\mathrm{a}} C.K.M. Lee ^(a){ }^{\mathrm{a}} 、Yaqiong Lv^(b**)\mathrm{Lv}^{\mathrm{b} *} 、K.K.H. Ng^(a)\mathrm{Ng}^{\mathrm{a}} 、William Ho^(c)\mathrm{Ho}^{\mathrm{c}} 和 K.L. Choy ^(a){ }^{\mathrm{a}}^(a){ }^{a} Department of Industrial and Systems Engineering, The Hong Kong Polytechnic University, Hong Kong, P.R.China; ^(b){ }^{b} School of Logistics Engineering, Wuhan University of Technology, Wuhan, P.R.China; ^(c){ }^{c} Department of Management and Marketing, University ^(a){ }^{a} 香港理工大学工业与系统工程系，香港，中国； ^(b){ }^{b} 武汉理工大学物流工程学院，武汉，中国； ^(c){ }^{c} 香港大学管理与市场营销系，香港，中国；香港大学管理与市场营销系，香港，中国；香港大学管理与市场营销系，香港，中国；香港大学管理与市场营销系，香港，中国。
of Melbourne, Melbourne, Australia 澳大利亚墨尔本
(Received 23 February 2017; accepted 26 September 2017) (2017年2月23日收到；2017年9月26日接受）
Abstract 摘要
Warehouse operations need to change due to the increasing complexity and variety of customer orders. The demand for real-time data and contextual information is requried because of the highly customised orders, which tend to be of small batch size but with high variety. Since the orders frequently change according to customer requirements, the synchronisation of purchase orders to support production to ensure on-time order fulfilment is of high importance. However, the inefficient and inaccurate order picking process has adverse effects on the order fulfilment. The objective of this paper is to propose an Internet of things (IoT)-based warehouse management system with an advanced data analytical approach using computational intelligence techniques to enable smart logistics for Industry 4.0. Based on the data collected from a case company, the proposed IoT-based WMS shows that the warehouse productivity, picking accuracy and efficiency can be improved and it is robust to order variability. 由于客户订单的复杂性和多样性与日俱增，仓库运营也需要做出改变。由于高度定制化的订单往往批量小但种类多，因此需要实时数据和上下文信息。由于订单经常会根据客户的要求发生变化，因此同步采购订单以支持生产，确保准时完成订单就显得尤为重要。然而，低效和不准确的订单分拣流程会对订单执行产生不利影响。本文旨在提出一种基于物联网（IoT）的仓库管理系统，该系统采用先进的数据分析方法，利用计算智能技术实现工业 4.0 的智能物流。根据从一家案例公司收集到的数据，所提出的基于物联网的仓库管理系统表明，仓库生产率、分拣准确性和效率都可以得到提高，并且对订单的变化具有鲁棒性。
Keywords: Internet of things; warehouse management system; low-volume, high-mix; Industry 4.0; smart logistics 关键词物联网；仓库管理系统；小批量、多品种；工业 4.0；智能物流
1. Introduction 1.导言
To enhance productivity and cope with the changing needs of customers, product design, production, packaging and distribution accordingly. In 2011, a new concept, Industry 4.0, was introduced in Germany. In the context of Industry 4.0, the future logistics on how physical objects are transported, handled, stored, supplied, realised and used across the world can be reshaped by Physical Internet so as for improvements in logistics efficiency and sustainability (Montreuil 2011) and cyber-physical systems and Internet of things (IoT) make it possible that industry-relevant items like materials, sensors, machines and products or in terms of Physical Internet containers are all connected and communicate with one another. All the connected items can be tracked and monitored so as to allow manufacturers to know the patterns and performance. With the decentralised intelligent decision-making, Industry 4.0 can be described as the increasing digitisation and automation of the manufacturing environment as well as the creation of a digital value chain that enables communication between products, the environment and business partners (Lasi et al. 2014). Industry 4.0 represents the coming fourth industrial revolution by adopting IoT, information and services for the next production paradigm. Decentralised intelligence helps create intelligent networks and optimise independent processes, with the interaction of the real and virtual worlds representing a crucial new milestone in industry development. Industry 4.0 represents a paradigm shift from ‘centralised’ to ‘decentralised’ production - made possible by technological advances, which constitute a reversal of conventional production process logic. Industrial production machinery no longer simply ‘process’ product, but the product communicates with the machinery to tell it exactly what to do (Wang et al. 2016). Industry 4.0 introduces an embedded system with latest production technologies. Smart production processes pave the way to a new technological age, which will radically transform industry and production value chains and business models. 为了提高生产率，应对客户、产品设计、生产、包装和分销不断变化的需求。2011 年，德国提出了一个新概念--工业 4.0。在 "工业 4.0 "的背景下，物理互联网将重塑未来物流，即物理物品在全球范围内的运输、处理、存储、供应、实现和使用方式，从而提高物流效率和可持续性（Montreuil，2011 年）。网络物理系统和物联网（IoT）使材料、传感器、机器和产品等工业相关物品或物理互联网容器相互连接和通信成为可能。所有联网的物品都可以被跟踪和监控，以便制造商了解其模式和性能。通过分散的智能决策，工业 4.0 可以被描述为制造环境的日益数字化和自动化，以及数字价值链的创建，从而实现产品、环境和业务合作伙伴之间的沟通（Lasi 等人，2014 年）。工业 4.0 代表着即将到来的第四次工业革命，它采用物联网、信息和服务来实现下一个生产范式。分散式智能有助于创建智能网络和优化独立流程，现实世界与虚拟世界的互动是工业发展的一个重要新里程碑。工业 4.0 代表着从 "集中式 "生产到 "分散式 "生产的范式转变--技术进步使之成为可能，颠覆了传统的生产流程逻辑。 工业生产机械不再只是简单地 "加工 "产品，而是产品与机械进行交流，告诉机械具体要做什么（Wang 等人，2016 年）。工业 4.0 引入了具有最新生产技术的嵌入式系统。智能生产流程为新技术时代铺平了道路，将从根本上改变工业、生产价值链和商业模式。
Apart from production, inbound logistics and outbound logistics pay an important role in fulfilling customer orders. The role of the warehouse has changed dramatically due to the complexity and variety of customer orders, the demand for real-time information and data accuracy. Therefore, it raises the problem that the traditional manual operation leads to low warehouse operation efficiency and is no longer responsive to the customer requirements. Among all the operations in the warehouse, research studies found that the order picking process can account for 50-55%50-55 \% of total operating expense (Frazelle and Frazelle 2002; de Koster, Le-Duc, and Roodbergen 2007). 除生产外，入库物流和出库物流在完成客户订单方面也发挥着重要作用。由于客户订单的复杂性和多样性，以及对实时信息和数据准确性的要求，仓库的角色发生了巨大变化。因此，传统的手工操作导致仓库运作效率低下，不再能满足客户的要求，这就提出了一个问题。研究发现，在仓库的所有操作中，订单拣选流程占总运营费用的 50-55%50-55 \% （Frazelle 和 Frazelle，2002 年；de Koster、Le-Duc 和 Roodbergen，2007 年）。
The contemporary warehouse management system (WMS) used by manufacturers is required to support the changes in production orders and to enhance the efficiency of the warehouse operation (Lee, Cao, and Ng 2017). Generally, WMS is always associated with auto-ID data capture technology in order to improve the inventory control and minimise the manual operation. The purpose of this research is to design and evaluate the effectiveness of the IoT-based Warehouse Inventory Management System for the low-volume, high-product mix situations faced by manufacturers, so as to achieve better performance of the receiving, storage and picking activities in the warehouse. 制造商使用的当代仓库管理系统（WMS）需要支持生产订单的变化，并提高仓库运作的效率（Lee、Cao 和 Ng，2017 年）。一般来说，WMS 总是与自动识别数据采集技术相结合，以改善库存控制并最大限度地减少人工操作。本研究的目的是针对制造商面临的小批量、高产品组合情况，设计基于物联网的仓库库存管理系统并评估其有效性，从而使仓库中的收货、存储和拣选活动取得更好的绩效。
Furthermore, the order picking process is the major bottleneck of the warehouse operation. Therefore, we propose WMS integrated with the fuzzy clustering technique in order to suggest the most suitable order picking method and to enhance the efficiency of the order picking process. Through the proposed WMS system, the warehouse activities, including the receiving, storage and order picking, can be managed and improved. 此外，订单拣选过程是仓库运作的主要瓶颈。因此，我们提出了与模糊聚类技术相结合的仓库管理系统，以建议最合适的订单拣选方法，提高订单拣选流程的效率。通过建议的 WMS 系统，可以管理和改进仓库活动，包括收货、存储和订单拣选。
In view of difficulties described above, the manual operation should be replaced by an advanced WMS. The functionality of the WMS, such as the order picking method, is proposed. An IoT device can provide the pickers’ work location information on the items, such that the efficiency of the order picking process can be enhanced. On the other hand, the high workload of the workers is the core problem. In the manual operation, it is very common that the worker places the product randomly and the picking process relies on worker’s memory and experience. Therefore, the operation is very time-consuming, and the workload of the worker is relatively a higher compared with a highly automated warehouse. As a result, the morale of the worker will be reduced, leading to a high turnover rate. 鉴于上述困难，应以先进的 WMS 系统取代手工操作。本文提出了 WMS 的功能，如订单分拣方法。物联网设备可以提供拣货员的物品工作位置信息，从而提高订单拣货流程的效率。另一方面，工人工作量大是核心问题。在手工操作中，工人随意摆放产品的情况非常普遍，拣选过程依赖于工人的记忆和经验。因此，操作非常耗时，与高度自动化的仓库相比，工人的工作量相对较大。因此，工人的士气会降低，导致高离职率。
Because of the above problems, an IoT-based WMS is proposed in order to minimise the warehouse operation process. The reduction of unnecessary processes can reduce the workload of the picker by applying the IoT-based technology in the receiving process instead of a manual paper record of the inventory. Therefore, it can help to improve the efficiency of the warehouse operation and increase the job satisfaction of the workers. 鉴于上述问题，我们提出了基于物联网的 WMS 系统，以最大限度地减少仓库操作流程。通过在收货过程中应用基于物联网的技术而不是手工纸质库存记录，可以减少不必要的流程，从而减轻拣货员的工作量。因此，这有助于提高仓库运作的效率，增加工人的工作满意度。
In this paper, Section 2 lists a current research review for warehouse management in coping with low-volume, highproduct mix with IoT technology. Section 3 outlines the whole framework of the proposed system and a case study is presented in Section 4 to validate our proposed system. The last section gives the conclusions and lists the limitations and future work. 在本文中，第 2 节列出了当前利用物联网技术应对小批量、多产品组合的仓库管理的研究综述。第 3 节概述了拟议系统的整体框架，第 4 节介绍了一个案例研究，以验证我们的拟议系统。最后一节给出了结论，并列出了局限性和未来工作。
2. Literature review 2.文献综述
2.1 Challenge of warehouse operation in the era of industry 4.0 2.1 工业 4.0 时代仓库运营面临的挑战
Inventory accuracy, space utilisation, process management and picking optimisation are the major challenge in modern warehouse management (Richards 2014). An agile supply chain strategy becomes a necessity in a supply chain network. In order to maintain smooth inbound and outbound logistics, there is a need to enhance the flexibility in the changing environment and reduce the total cycle time of a supply chain system. The cyber-physical systems (CPS) network becomes a mediator to connect people, objects and physical processes in a warehouse operation over the IoT, a wireless network (Culler and Long 2016). The emergence of the CPS network fosters responsiveness and flexibility in WMS (Leitão, Colombo, and Karnouskos 2016). Evolution from the traditional WMS to CPS-WMS requires the integration of technological and administrative innovations, and this becomes the major challenge for the design of WMS. These include proper selection of CPS technologies, ambient intelligence, timely information flow and agility (Reaidy, Gunasekaran, and Spalanzani 2015). 库存准确性、空间利用率、流程管理和分拣优化是现代仓储管理的主要挑战（Richards，2014 年）。灵活的供应链战略成为供应链网络的必要条件。为了保持顺畅的入库和出库物流，需要在不断变化的环境中提高灵活性，缩短供应链系统的总周期时间。网络物理系统（CPS）网络通过无线网络物联网（Culler 和 Long，2016 年）成为仓库运作中连接人、物和物理流程的中介。CPS 网络的出现提高了 WMS 的响应速度和灵活性（Leitão、Colombo 和 Karnouskos，2016 年）。从传统的 WMS 向 CPS-WMS 演进需要整合技术和管理创新，这成为 WMS 设计的主要挑战。这包括正确选择 CPS 技术、环境智能、及时的信息流和敏捷性（Reaidy、Gunasekaran 和 Spalanzani，2015 年）。
2.1.1 CPS technology in WMS 2.1.1 仓库管理系统中的 CPS 技术
The implementation of the CPS-WMS assists in the establishment of cooperation between human, intelligent machines and robots, transparent in the performance of a smart WMS (Posada et al. 2015). Reviewing the nine pillars of technology in supporting the development of CPS-WMS, the common technologies includes radio frequency identification/ near-field communication (RFID/NFC), wireless sensor and actuator networks (WSANs), IoT and Cloud computing (Qiu et al. 2015; Wan et al. 2016). CPS network synergises the growth of big data analytics with these separated keyenabling technologies to provide insight towards greater value proposition, analytical powers and decision-making process (Waller and Fawcett 2013). CPS-WMS 的实施有助于在人类、智能机器和机器人之间建立合作关系，使智能 WMS 的性能透明化（Posada 等人，2015 年）。回顾支持 CPS-WMS 发展的九大技术支柱，常见技术包括射频识别/近场通信（RFID/NFC）、无线传感器和执行器网络（WSAN）、物联网和云计算（Qiu 等，2015 年；Wan 等，2016 年）。CPS 网络将大数据分析的发展与这些分离的关键技术协同起来，为更大的价值主张、分析能力和决策过程提供洞察力（Waller 和 Fawcett，2013 年）。
2.1.2 Ambient intelligence 2.1.2 环境智能
The availability of CPS technology in WMS helps to facilitate the traceability and transparency in warehouse operations via the use of ambient intelligence (Olaru and Gratie 2011). The heterogeneity and composability of the ambient 在仓库管理系统中采用 CPS 技术有助于通过使用环境智能促进仓库运作的可追溯性和透明度（Olaru 和 Gratie，2011 年）。环境智能的异质性和可组合性有助于提高仓库操作的可追溯性和透明度（Olaru 和 Gratie，2011 年）。
intelligence system allow the system to detect the activities and interactions of operation within the warehouse (Atmojo et al. 2015). In addition, the system contains multiple synchronous decisions. If an improper software framework is developed, a deadlock condition may arise during the running of the system (Atmojo et al. 2015). 智能系统允许系统检测仓库内的操作活动和互动（Atmojo 等人，2015 年）。此外，系统还包含多个同步决策。如果开发的软件框架不当，系统运行过程中可能会出现死锁情况（Atmojo 等人，2015 年）。
2.1.3 Real-time information sharing 2.1.3 实时信息共享
With the purpose of remaining a high level of agility in CPS-WMS, the CPS network requires real-time information monitoring and visibility of the CPS system among all the operations and activities within the WMS (Reaidy, Gunasekaran, and Spalanzani 2015). The advanced connectivity of data acquisition between the physical operations and visual system is of the essence (Lee, Bagheri, and Kao 2015). Real-time information sharing enables right decision support and coping with the changing requirement from customers. 为了保持 CPS-WMS 的高度灵活性，CPS 网络需要对 CPS 系统进行实时信息监控，并确保 WMS 内所有操作和活动的可见性（Reaidy、Gunasekaran 和 Spalanzani，2015 年）。物理操作和可视系统之间先进的数据采集连接至关重要（Lee、Bagheri 和 Kao，2015 年）。实时信息共享可提供正确的决策支持，并应对客户不断变化的需求。
2.2 Latest research development of IoT for WMS 2.2 物联网在仓库管理系统中的最新研究进展
WSANs create transparency and value in WMS with more vigorous and sophisticated support decision-making. IoT becomes the essential element in CPS-WMS, which enhances the visibility and real-time taking in management through WSANs. Deploying automated data acquirement enables communication between warehouse operations in the cloud platform or big data infrastructure (Tracey and Sreenan 2013). Intelligent WMS enhances the tally process, simplifies the operations and increases the degree of automated WMS (Ding 2013). WSAN 为 WMS 创造了透明度和价值，为决策提供更有力、更复杂的支持。物联网成为 CPS-WMS 的基本要素，通过 WSAN 提高了管理的可视性和实时性。部署自动数据采集系统可在云平台或大数据基础设施中实现仓库操作之间的通信（Tracey 和 Sreenan，2013 年）。智能仓库管理系统（WMS）增强了理货流程，简化了操作，提高了仓库管理系统的自动化程度（Ding，2013 年）。
Application of IoT-based WMS has become popular. RFID technology is widely adopted in warehouse management, as the technology allows trace and track and identification of specified objects. Chow et al. (2006) proposed a RFIDbased WMS for the retrieval and matching process of customer orders to enhance the throughput of the warehouse and provide an accurate inventory monitoring system. Poon et al. (2009) utilised RFID-based order-picking operations to reduce the likelihood of operational errors. Besides, the integration of RFID-based WMS and Enterprise Resource Planning encourage the development of Event-driven Process Chains (EPC) in business process management (Liu, Jeng, and Chang 2008). Furthermore, WSNs is another complementary research approach in assisting information extraction on the conditions of objects. However, it is challenging in data acquisition, distribution and mining (Wang et al. 2014). In order to obtain complete logistics order tracking in the tobacco supply chain, the tracking and delivery in in-bound and out-bound logistics was reviewed by Jiang and Su (2013) employing Global Positioning System (GPS), Geographic Information System (GIS) and General Packet Radio Service (GPRS). Yang (2012) proposed a location-based system for forklifts in order to monitor the logistics activities in an intelligent warehouse. 基于物联网的仓库管理系统的应用已变得十分普遍。射频识别（RFID）技术在仓库管理中被广泛采用，因为该技术可以对指定物品进行追踪和识别。Chow 等人（2006 年）提出了一种基于 RFID 的 WMS 系统，用于客户订单的检索和匹配过程，以提高仓库的吞吐量，并提供准确的库存监控系统。Poon 等人（2009 年）利用基于射频识别（RFID）的订单拣选操作来减少操作失误的可能性。此外，基于 RFID 的 WMS 与企业资源规划的整合促进了业务流程管理中事件驱动流程链（EPC）的发展（Liu、Jeng 和 Chang，2008 年）。此外，WSN 是另一种辅助研究方法，可协助提取有关物体状况的信息。然而，它在数据采集、分发和挖掘方面具有挑战性（Wang 等人，2014 年）。为了在烟草供应链中获得完整的物流订单跟踪，Jiang 和 Su（2013 年）采用全球定位系统（GPS）、地理信息系统（GIS）和通用分组无线服务（GPRS）对进出物流的跟踪和交付进行了研究。Yang（2012）提出了一种基于位置的叉车系统，以监控智能仓库中的物流活动。
The deployment of IoT facilitates the development of the automated warehouse. Kim and Sohn (2009) introduced a control system for managing industrial machines, resources and products via IoT in an information technology infrastructure. Alyahya, Wang, and Bennett (2016) further studied the feasibility of an RFID-enabled automated storage and retrieval system without manual intervention. Bajic (2009) presented a platform with IoT and the ambient network between the product, process, environment and users for agent warehousing management. The proposed agent-based WMS, which allows remote action invocation, becomes the service-based control point in providing a high level of manageable capability and the enhancement of operational efficiency. Finally, process control in WMS comes in to the CPS network era, which further enhances the control level from automated WMS to virtual synchrony of physical objects. This perspective allows WSANs-based communication in an autonomous CPS-WMS. 物联网的部署促进了自动化仓库的发展。Kim 和 Sohn（2009 年）在信息技术基础设施中引入了一种通过物联网管理工业机器、资源和产品的控制系统。Alyahya、Wang 和 Bennett（2016 年）进一步研究了无需人工干预的 RFID 自动化存储和检索系统的可行性。Bajic（2009）提出了一个物联网平台以及产品、流程、环境和用户之间的环境网络，用于代理仓储管理。所提出的基于代理的 WMS 允许远程操作调用，成为基于服务的控制点，提供高水平的可管理能力并提高运营效率。最后，WMS 的流程控制进入了 CPS 网络时代，进一步提升了从自动化 WMS 到实物虚拟同步的控制水平。从这个角度看，在自主的 CPS-WMS 中，基于 WSAN 的通信是可行的。
Although there are more findings of applying the IoT concept in WMS to enable autonomous characteristics for CPS, there has been sparse research illustrating the practical and operational level of WMS, such as order picking using computational intelligence with IoT to enable smart logistics. 虽然在 WMS 中应用物联网概念实现 CPS 自主特性的研究成果较多，但在 WMS 的实际操作层面，如利用计算智能和物联网实现智能物流的订单分拣方面的研究却很少。
3. Framework of proposed WMS 3.拟议的 WMS 框架
The purpose of this research is to design and evaluate the effectiveness of an IoT-based WMS for a low-volume, highproduct mix scenario. Due to the complexity of IoT data synchronisation in this WMS, the status tracking and connection of the cyber-physical system is extremely important so as to maintain data consistency. This is why the proposed IoT-based WMS system is highly desirable. In this section, the workflow of the low-volume, high-product mix WMS is proposed, and the whole framework is provided and embedded with appropriate techniques to handle different problems at each stage. 本研究的目的是设计和评估基于物联网的 WMS 系统在低产量、高产品组合情况下的有效性。由于该 WMS 系统中物联网数据同步的复杂性，网络物理系统的状态跟踪和连接对于保持数据一致性极为重要。因此，基于物联网的 WMS 系统非常值得推荐。本节提出了小批量、多产品组合 WMS 的工作流程，并为整个框架提供和嵌入了适当的技术，以处理每个阶段的不同问题。
Warehouse activities include the inbound area activities and outbound areas activities, like receiving, storage, quality inspection, picking and shipping. However, in the low-volume, high-mix industrial environment, other than the inbound and outbound areas, the internal processing is another very crucial part in warehouse management as this particular warehouse configuration allows a higher flexibility of purchase order (PO) change. 仓库活动包括进货区活动和出货区活动，如收货、储存、质量检验、分拣和发货。然而，在小批量、多品种的工业环境中，除了入库和出库区域外，内部处理是仓库管理中另一个非常关键的部分，因为这种特殊的仓库配置允许采购订单（PO）变更具有更高的灵活性。
Figure 1 illustrates the warehouse activities with the typical workflow of a low-volume, high-mix scenario. When a new order comes in, the Bill of Material is generated based on the required quantity of different raw materials such that the production process can start on time, without any delay. If the required materials are available, the picking activities proceed from like inbound store or sub-store. After checking the current inventory, if the quantity of the required materials is inadequate, a PO for the shortage is generated. It triggers the inbound logistics activities so that the outstanding amount of material is received afterwards. In between, the internal engine is responsible for changing the orders. If the customers call for pull-in, push-out or cancel of a certain order, the PO adjust to the different situations to update the corresponding information. The updated PO leads to inbound warehouse activities, which correspond to the new PO. 图 1 展示了小批量、多品种典型工作流程下的仓库活动。当收到新订单时，会根据不同原材料的所需数量生成物料清单，以便生产流程能及时启动，不会出现任何延误。如果所需材料齐备，拣选活动就会从类似的入库仓库或分库开始。检查当前库存后，如果所需材料数量不足，就会生成缺货 PO。这将触发入库物流活动，以便随后收到所欠的材料数量。在此期间，内部引擎负责更改订单。如果客户要求拉入、推出或取消某个订单，PO 就会根据不同的情况调整，更新相应的信息。更新后的 PO 将导致与新 PO 相对应的入库仓库活动。
3.2 IoT-based WMS 3.2 基于物联网的 WMS
In the highly customised and flexible low-volume, high-product mix industry, the involved raw parts and semi-finished goods are in small amounts with high variety. The information exchange and updating is a crucial problem in handling the new orders, while order changes always occur. The proposed IoT-based WMS fully utilises RFID technology and wireless sensors to track and trace the raw parts, semi-finished goods and finished goods. The embedded system helps collect all the information changes and updates of the warehouse activities. With IoT technology, the incoming parts and activities are all controlled, and the inconsistency due to order change or updating can be automatically handled and solved by the proposed system. 在高度定制和灵活的小批量、多产品组合行业中，涉及的原材料和半成品数量少、种类多。在处理新订单时，信息交换和更新是一个关键问题，而订单变更总是时有发生。拟议的基于物联网的 WMS 系统充分利用 RFID 技术和无线传感器来跟踪和追溯原材料、半成品和成品。嵌入式系统有助于收集仓库活动的所有信息变更和更新。利用物联网技术，入库部件和活动都能得到控制，而且由于订单变更或更新而造成的不一致问题也能由拟议的系统自动处理和解决。
Figure 2 shows the framework of the proposed IoT-based WMS: raw material, semi-finished goods and finished goods are stored in the warehouse, sub-store, or are pending for delivery in the distribution centre. In the IoT environment, all the parts are labelled with an RFID tag. The parts or products are identified from the RF reader antenna and then information is transmitted to the radio identification reader and subsequently through the RFID middleware to the EPC information server. The host application then integrates applications according to different needs (Lv et al. 2012, 图 2 显示了拟议的基于物联网的 WMS 系统框架：原材料、半成品和成品存储在仓库、子仓库中，或在配送中心等待交付。在物联网环境中，所有部件都贴有 RFID 标签。部件或产品通过射频阅读器天线进行识别，然后将信息传输到射频识别阅读器，再通过射频识别中间件传输到 EPC 信息服务器。然后，主机应用程序根据不同需求整合应用程序（Lv 等人，2012 年）、
Figure 2. The framework of proposed IoT-based warehouse management system. 图 2.拟议的基于物联网的仓库管理系统框架。
Figure 3. The function of internal order-change handle engine. 图 3.内部订单更改处理引擎的功能。
2013). Since this proposed system is designed for a low-volume high-mix scenario, which is the typical situation faced by the manufacturers in the Industry 4.0 era, the customisation of orders requires high flexibility of information updating. That is why the collected information by RFID can be mortified or deleted by authorised staff via mobile apps any place at any time. To synchronise and optimise the inventory, the data and information are inputted to an intelligent inventory management engine to handle order change and picking problems, among which data clustering and some machine learning methods as well as fuzzy inference system are applied for information processing in decision-support. The output is transmitted back to the host application and shares the results with the mobile apps. Hence, the staff involved in this IoT-based WMS can receive the corresponding action information. 2013).由于所提议的系统是针对小批量、多品种的情况而设计的，而这正是工业 4.0 时代制造商所面临的典型情况，因此订单的定制需要高度灵活的信息更新。因此，经授权的员工可随时随地通过移动应用程序对 RFID 收集的信息进行修改或删除。为了同步和优化库存，数据和信息被输入到智能库存管理引擎，以处理订单变更和分拣问题，其中数据聚类和一些机器学习方法以及模糊推理系统被应用于决策支持的信息处理。输出结果会传回主机应用程序，并与移动应用程序共享结果。因此，参与这个基于物联网的 WMS 系统的工作人员可以收到相应的行动信息。
As specified in Section 3.1, the inbound logistics for receiving goods can be easily settled by IoT technology. The internal engine focuses on handling order change, as shown in Figure 3. The input is the change requested from customers on existing orders, inclusive of pull-in, push-out and cancellation. In the low-volume, high-mix scenario, most of the parts under a certain PO include common parts, and because of the Minimal Order Quantity one PO may serve for several job orders. Therefore, the internal engine will build up an efficient and effective logical, rule-based engine to provide the correct action, which avoids affecting other job schedules. 如第 3.1 节所述，接收货物的入站物流可以通过物联网技术轻松解决。内部引擎主要处理订单变更，如图 3 所示。输入是客户对现有订单提出的变更要求，包括拉入、推出和取消。在小批量、多品种的情况下，某个订单下的大部分零件都是通用零件，而且由于最小订单量的存在，一个订单可能会同时服务于多个工作订单。因此，内部引擎将建立一个高效的、基于规则的逻辑引擎，提供正确的操作，避免影响其他作业计划。
After all, if all the POs are confirmed with sufficient inventory, another significant operation of the warehouse management is order picking. Picking is more complex and difficult than the receiving processes, and the proposed system 毕竟，如果所有 PO 都已确认有足够的库存，那么仓库管理的另一项重要操作就是订单拣选。与收货流程相比，拣货流程更为复杂和困难。
integrated with the fuzzy logic technique suggests the most suitable order picking method to enhance the efficiency of operation. One of the advantages of the fuzzy logic model over other approaches is that it is easier for the end-user to understand through its linguistic fuzzy terms, fuzzy values and logical reasoning process. For the qualitative attributes like configurability, the outstanding easy-to-understand feature of our model makes it useful for non-numerical or insufficient input-data assessment feasible to meet real-life needs. Its second advantage is the ability to adopt end-user’s domain knowledge or business logic into the measurement process through refining or customising its fuzzy systems. Figure 4 shows the picking process with fuzzy logic technique. 与模糊逻辑技术相结合，提出了最合适的订单拣选方法，以提高操作效率。与其他方法相比，模糊逻辑模型的优势之一是，通过语言模糊术语、模糊值和逻辑推理过程，最终用户更容易理解。就可配置性等定性属性而言，我们的模型突出的易懂特点使其适用于非数字或输入数据不足的评估，从而满足现实生活的需要。它的第二个优势是能够通过完善或定制模糊系统，将最终用户的领域知识或业务逻辑纳入测量流程。图 4 显示了采用模糊逻辑技术的采摘流程。
When the warehouse operators receive the goods from the production department, information on the goods such as SKU number, PO number, customer details, quantity and the location is captured by IoT technology in the data collection module. Such information is taken into account to generate the best order picking method. In this order picking module, the fuzzy logic engine is involved. In this module, the fuzzy logic theory is applied in order to assess the most appropriate method of order picking to improve the efficiency of the order picking process. Fuzzification is the first step in the fuzzy logic engine. After the data are collected by the RFID, the input data are converted into the fuzzy set and the characteristic is mainly determined by the membership function using the formulation: 当仓库操作员从生产部门收到货物时，数据收集模块中的物联网技术会捕捉到货物的信息，如 SKU 编号、PO 编号、客户详细信息、数量和位置。这些信息将作为生成最佳订单分拣方法的参考。在订单拣选模块中，涉及到模糊逻辑引擎。在该模块中，应用了模糊逻辑理论，以评估最合适的订单分拣方法，从而提高订单分拣流程的效率。模糊化是模糊逻辑引擎的第一步。通过 RFID 采集数据后，输入数据被转换为模糊集，其特征主要由使用公式的成员函数决定：
where i=1,2,dots,ci=1,2, \ldots, c and k=1,2,dots,nk=1,2, \ldots, n. 其中， i=1,2,dots,ci=1,2, \ldots, c 和 k=1,2,dots,nk=1,2, \ldots, n 。
In the engine, the working principle of the inference process is to transfer the input fuzzy set into the fuzzy inference engine, where the process involves rule block formation and rule composition. The final step of the fuzzy logic is defuzzification. In defuzzification, Graded Mean Integration Representation is adopted in order to calculate the results. The GMIR can be described as follows: 在该引擎中，推理过程的工作原理是将输入的模糊集转入模糊推理引擎，这一过程包括规则块形成和规则构成。模糊逻辑的最后一步是去模糊化。在去模糊化过程中，采用分级平均积分表示法来计算结果。分级平均积分表示法可描述如下：
suppose L^(-1)L^{-1} and R^(-1)R^{-1} are inverse functions of functions LL and RR, respectively, and the graded mean h-level value of generalised fuzzy number A=(c,a,b,d;w)_(LR)A=(c, a, b, d ; w)_{L R} is h[L^(-1)(h)+R^(-1)(h)]//2h\left[L^{-1}(h)+R^{-1}(h)\right] / 2, then the graded mean integration representation of generalised fuzzy number based on the integral value of graded mean hh-levels is: 假设 L^(-1)L^{-1} 和 R^(-1)R^{-1} 分别是函数 LL 和 RR 的反函数，且广义模糊数 A=(c,a,b,d;w)_(LR)A=(c, a, b, d ; w)_{L R} 的分级平均 h 级值为 h[L^(-1)(h)+R^(-1)(h)]//2h\left[L^{-1}(h)+R^{-1}(h)\right] / 2 ，则基于分级平均 hh 级积分值的广义模糊数分级平均积分表示为
Figure 4. Picking processes with fuzzy logic technique. 图 4.采用模糊逻辑技术的采摘流程。
P(A)=(int_(0)^(w)h((L^(-1)(h)+R^(-1)(h))/(2))dh)/(int_(0)^(w)h(d)h)P(A)=\frac{\int_{0}^{w} h\left(\frac{L^{-1}(h)+R^{-1}(h)}{2}\right) \mathrm{d} h}{\int_{0}^{w} h \mathrm{~d} h}
where hh is between 0 and w,0 < w <= 1w, 0<w \leq 1. 其中 hh 介于 0 和 w,0 < w <= 1w, 0<w \leq 1 之间。
Through this process, the crisp values can be generated so as to assess the order picking method for the picker’s operation in the picking activity, with respective actions such as strict order picking, batch picking, sequential zone picking, batch zone picking, wave picking. Table 1 summarises the order picking a policy and analyses the advantages and disadvantages of each policy (Ii 2000). 通过这一过程，可以生成清晰的数值，从而评估拣选活动中拣选工操作的订单拣选方法，分别有严格订单拣选、批量拣选、顺序分区拣选、批量分区拣选、波浪拣选等操作。表 1 总结了订单拣选策略，并分析了每种策略的优缺点（Ii 2000）。
Table 1. Comparison of order picking policy. 表 1.订单分拣策略比较。
Order picking policy 订单拣选政策
Description 说明
Advantage 优势
Disadvantage 劣势
Strict order picking 严格的拣选顺序
Each worker handles one order at one time. In other words, a worker will go to one picking station and finish one order 每个工人一次处理一个订单。换句话说，一名工人将前往一个分拣站，完成一份订单
为工人提供简单明了的订单分拣方法 允许工人直接检查错误 无需重新处理货物，如分类
Simple and clear order picking approach for the worker
Allows direct error checking by the worker
Does not require re-handling of the goods such as sorting
Simple and clear order picking approach for the worker
Allows direct error checking by the worker
Does not require re-handling of the goods such as sorting| Simple and clear order picking approach for the worker |
| :--- |
| Allows direct error checking by the worker |
| Does not require re-handling of the goods such as sorting |
Lower efficiency for an order with a lot of single items 单品较多的订单效率较低
Batch picking 批量采摘
Batch Order picking means that each worker will pick more than one order at one time. In other words, the worker will go to one picking station and finish several orders 批量订单分拣是指每个工人一次分拣多个订单。换句话说，工人将前往一个分拣站，完成多个订单的分拣。
Higher efficiency Less travel time per item 效率更高 每件物品的运输时间更短
比严格的订单分拣更复杂。需要对每批订单进行分拣，分拣需要空间 更多潜在错误
More complicated than the strict order picking. Sorting for each batch order is required, and space for sorting is required
More potential error
More complicated than the strict order picking. Sorting for each batch order is required, and space for sorting is required
More potential error| More complicated than the strict order picking. Sorting for each batch order is required, and space for sorting is required |
| :--- |
| More potential error |
Sequential zone picking 顺序分区
The worker will pick one order at a time, and the picking sequence is from zone to zone. The worker will pick the goods from their zone and then pass the picking list to another worker. The next worker also will pick the goods from their own zone and then pass it again until the picking list is finished and then go through all the zones 工人每次分拣一个订单，分拣顺序是从一个区域到另一个区域。工人从自己的区域拣选货物，然后将拣选清单交给另一名工人。下一位工人也将从自己的分区拣选货物，然后再次传递，直到拣选清单上的货物拣选完毕，然后走完所有分区。
Suitable for a large distribution centre No sorting is required Increase responsibility of picker and house keeping 适合大型配送中心 无需分拣 增加拣货员和内务管理的责任
Difficult to define the zone and zone capacity Imbalanced workload in picking zone 难以确定分区和分区容量 采摘区工作量不平衡
Batch zone picking 批量区域拣选
Orders are picked and put on the conveyor belt and sent to other zones. Sorting is conducted in the final area 订单被拣选并放到传送带上，然后送往其他区域。在最后区域进行分拣
Volume picking of single or several items is allowed 允许批量拣选单个或多个物品
Loss of order integrity Error of picking and sorting increases the chance of errors Imbalanced workload in picking zone 订单完整性的损失 分拣和分类错误增加了出错的几率 分拣区工作量不平衡
Wave picking 选波
Wave picking occurs when the worker picks a batch items requiring long completion times. In this situation, the worker will finish the first wave and then start to pick the second wave. The wave picking is not finished until all the waves are picked 当工人分拣一批需要较长完成时间的物品时，就会出现分波分拣。在这种情况下，工人将完成第一波拣选，然后开始拣选第二波。直到所有波次都拣选完毕，波次拣选才算完成。
Maybe higher efficiency compared with batch picking in a large distribution warehouse 与大型配送仓库的分批拣选相比，效率可能更高
Loss of order integrity Error of picking and sorting increases the chance of error Imbalance workload in picking zone More time for order consolidation 订单完整性的损失 分拣和分类的错误增加了出错的几率 分拣区工作量不平衡 订单整合时间增加
Order picking policy Description Advantage Disadvantage
Strict order picking Each worker handles one order at one time. In other words, a worker will go to one picking station and finish one order "Simple and clear order picking approach for the worker
Allows direct error checking by the worker
Does not require re-handling of the goods such as sorting" Lower efficiency for an order with a lot of single items
Batch picking Batch Order picking means that each worker will pick more than one order at one time. In other words, the worker will go to one picking station and finish several orders Higher efficiency Less travel time per item "More complicated than the strict order picking. Sorting for each batch order is required, and space for sorting is required
More potential error"
Sequential zone picking The worker will pick one order at a time, and the picking sequence is from zone to zone. The worker will pick the goods from their zone and then pass the picking list to another worker. The next worker also will pick the goods from their own zone and then pass it again until the picking list is finished and then go through all the zones Suitable for a large distribution centre No sorting is required Increase responsibility of picker and house keeping Difficult to define the zone and zone capacity Imbalanced workload in picking zone
Batch zone picking Orders are picked and put on the conveyor belt and sent to other zones. Sorting is conducted in the final area Volume picking of single or several items is allowed Loss of order integrity Error of picking and sorting increases the chance of errors Imbalanced workload in picking zone
Wave picking Wave picking occurs when the worker picks a batch items requiring long completion times. In this situation, the worker will finish the first wave and then start to pick the second wave. The wave picking is not finished until all the waves are picked Maybe higher efficiency compared with batch picking in a large distribution warehouse Loss of order integrity Error of picking and sorting increases the chance of error Imbalance workload in picking zone More time for order consolidation| Order picking policy | Description | Advantage | Disadvantage |
| :---: | :---: | :---: | :---: |
| Strict order picking | Each worker handles one order at one time. In other words, a worker will go to one picking station and finish one order | Simple and clear order picking approach for the worker <br> Allows direct error checking by the worker <br> Does not require re-handling of the goods such as sorting | Lower efficiency for an order with a lot of single items |
| Batch picking | Batch Order picking means that each worker will pick more than one order at one time. In other words, the worker will go to one picking station and finish several orders | Higher efficiency Less travel time per item | More complicated than the strict order picking. Sorting for each batch order is required, and space for sorting is required <br> More potential error |
| Sequential zone picking | The worker will pick one order at a time, and the picking sequence is from zone to zone. The worker will pick the goods from their zone and then pass the picking list to another worker. The next worker also will pick the goods from their own zone and then pass it again until the picking list is finished and then go through all the zones | Suitable for a large distribution centre No sorting is required Increase responsibility of picker and house keeping | Difficult to define the zone and zone capacity Imbalanced workload in picking zone |
| Batch zone picking | Orders are picked and put on the conveyor belt and sent to other zones. Sorting is conducted in the final area | Volume picking of single or several items is allowed | Loss of order integrity Error of picking and sorting increases the chance of errors Imbalanced workload in picking zone |
| Wave picking | Wave picking occurs when the worker picks a batch items requiring long completion times. In this situation, the worker will finish the first wave and then start to pick the second wave. The wave picking is not finished until all the waves are picked | Maybe higher efficiency compared with batch picking in a large distribution warehouse | Loss of order integrity Error of picking and sorting increases the chance of error Imbalance workload in picking zone More time for order consolidation |
This above depicts how the inbound area, internal engine and outbound area work is used to handle the whole order picking problems with the proposed IoT-based inventory management system. This system connects and communicates all the parts/goods, locations and workers together in real-time with the embedded intelligent warehouse management engine, which provides suggestions for corresponding actions in different scenarios. 以上描述了入库区、内部引擎和出库区的工作是如何通过拟议的基于物联网的库存管理系统来处理整个订单分拣问题的。该系统通过嵌入式智能仓库管理引擎实时连接和通信所有部件/货物、地点和工人，并在不同场景下提供相应的行动建议。
4. Case study 4.案例研究
In this section, a manufacturing company (with an alias name CCI) is discussed and studied. CCI is a reputable manufacturer of Box Build and Equipment Manufacturing, who focuses on High-Mix, Low-Volume Contract Manufacturing Services and Equipment Integration. Due to CCI’s high-mix low-volume mode, the manufacturing material components accumulate enormously in the warehouse. With the high complexity of raw materials and semi-finished goods, CCI encounters difficulties in maintaining the inventory at a reasonable level. 本节将讨论和研究一家制造公司（别名 CCI）。CCI 是一家声誉卓著的箱体制造和设备制造公司，专注于多品种、小批量的合同制造服务和设备集成。由于 CCI 采用多品种、小批量的模式，制造材料组件在仓库中堆积如山。由于原材料和半成品的高度复杂性，CCI 在将库存保持在合理水平方面遇到了困难。
CCI’s low-volume, high-mix manufacturing mode determines its order-driven material inventory management process. With IoT technology, the material flow is controlled, and at the same time, the information flow is transparent in the company, which allows the planners to respond quickly to any new situation. Compared with traditional forecasting and planning, IoT allows users to obtain more relevant contextual information on the environment through sensors, actuators and computation tools to ensure smart behaviour. Accurate forecasting methods can calculate the material requirement for mass production. However, CCI faces typical low-volume, high-mix situations and forecasting are not easy, with wrong forecasting always leading to high inventory. That is why CCI operates on an order-driven mode, which requires material real-time monitoring capability and prompt order-change handling capability. Under such circumstances, the proposed IoT-based WMS is introduced to CCI for monitoring a wide variety of materials and in handling different scenarios automatically. This section discusses how the proposed IoT-based WMS works for CCI. CCI 的小批量、多品种生产模式决定了其订单驱动的物料库存管理流程。通过物联网技术，物料流得到了控制，同时，信息流在公司内部实现了透明化，使计划人员能够对任何新情况做出快速反应。与传统的预测和计划相比，物联网可以让用户通过传感器、执行器和计算工具获得更多相关的环境信息，从而确保智能行为。准确的预测方法可以计算出大规模生产所需的材料。然而，CCI 面临的是典型的小批量、多品种情况，预测并非易事，错误的预测总是导致高库存。因此，CCI 采用订单驱动模式运行，这就要求具备物料实时监控能力和及时订单变更处理能力。在这种情况下，拟议的基于物联网的 WMS 被引入 CCI，用于监控各种材料并自动处理不同的情况。本节将讨论拟议的基于物联网的 WMS 如何在 CCI 中发挥作用。
4.1 Processing engine for PO synchronisation 4.1 PO 同步处理引擎
Figure 5 presents the workflow of CCI’s inbound logistics activities. Firstly, the goods are received and temporarily stored in the inbound area of the warehouse. Then, the workers do inspections such as checking the quality and counting the amount of the incoming goods. If there are defects or non-conformance of the purchase requirement, the goods 图 5 展示了 CCI 入库物流活动的工作流程。首先，接收货物并将其暂时存放在仓库的进货区。然后，工人对入库货物进行检查，如检查质量和清点数量。如果存在缺陷或不符合采购要求，则将货物退回。
Figure 5. Workflow of CCI’s inbound inventory processes. 图 5.CCI 入库流程的工作流程。
will be returned to the manufacturer. If the goods meet the requirement and meet the conformance level, the storage process will proceed. 将退回制造商。如果货物符合要求并达到合格水平，则继续执行存储程序。
In the storage process, the workers make use of RFID technology to record the information. In this process, there is a potential error such as recording the inventory higher or less than the actual inventory received. After the worker records the goods information, a check is needed as to whether the record is correct or not. If the record is correct, the storage process is complete. If the record is incorrect, the worker checks out the wrong record and deletes it. After the deletion process, the worker re-records the goods. 在存储过程中，工作人员利用 RFID 技术记录信息。在此过程中，可能会出现错误，如记录的库存高于或低于实际收到的库存。工人记录货物信息后，需要检查记录是否正确。如果记录正确，则存储过程完成。如果记录不正确，工人会检查出错误的记录并将其删除。删除过程结束后，工人重新记录货物信息。
Meanwhile, information flow is involved. Once the worker uses the technology record the goods information such as the storage location and goods quantity, all information passes through the wireless update in the WMS. Once the data are recorded in the data collection module, the data passes to the order-picking module as the input variable and the best order picking method is then generated. 同时，还涉及信息流。一旦工人使用该技术记录了货物信息，如存储位置和货物数量，所有信息都会通过无线更新传送到 WMS 系统中。数据收集模块记录数据后，数据作为输入变量传递给订单拣选模块，然后生成最佳订单拣选方法。
As mentioned in Section 3, work at this stage is quite straightforward with the incoming goods record. Figure 6 is the interface of the mobile apps, which is proposed for CCI to record and update the information of the various items. 如第 3 节所述，这一阶段的工作与进货记录相当直接。图 6 是移动应用程序的界面，建议 CCI 使用它来记录和更新各种物品的信息。
4.2 Internal processing engine for PO synchronisation 4.2 PO 同步的内部处理引擎
When CCI’s customers request certain projects are having a low-volume, high-mix requirement, CCI builds up many jobs under the project in the ERP system with corresponding information like item number, scheduled start date, required quantity, etc. Through CCI’s ERP system, the required items such as quantity on hand, order lead time and costs. as shown in Figure 7, can be checked. 当CCI的客户要求某些项目具有小批量、多品种的要求时，CCI会在ERP系统中的项目下建立许多工作，并提供相应的信息，如项目编号、计划开始日期、所需数量等。如图 7 所示，通过 CCI 的企业资源规划系统，可以检查所需的项目，如库存数量、订货提前期和成本。
Normally for a particular project, jobs will be generated at the same time to get a holistic understanding on any shortages. Once there is a shortage of a certain item, a PO will be generated. But in the low-volume, high-mix situation, items are always used across different jobs/projects. Due to different scheduled start dates, the need date for same item will be different for different jobs. That means one PO will serve multi-jobs. However due to the various lead time of items, the PO promised by the vendors cannot be confirmed at the specific times. CCI faces the problem that for a particular job, not all the PO, can arrive on time to meet the scheduled start date. If items on one PO arrive later than other PO, it will definitely lead to waiting. This is a crucial issue to bring up high inventory for CCI. Therefore, CCI wants to have an intelligent engine to help them solve such problems. If some PO is confirmed as late, is it possible to push out some existing POs without affecting other jobs? This engine is designed to point out possible push-out items to reduce inventory. Here, a simple example is depicted to explain what the engine does is shown in Figure 8. 通常情况下，对于一个特定项目，会同时生成作业，以便全面了解任何短缺情况。一旦某个项目出现短缺，就会生成采购单。但在小批量、多品种的情况下，物品总是被用于不同的工作/项目。由于计划开始日期不同，不同工作对同一物品的需求日期也会不同。这就意味着一个采购单可以服务于多个工作。然而，由于物品的交货期不同，供应商承诺的 PO 无法在特定时间得到确认。CCI 面临的问题是，对于某项工作，并非所有的 PO 都能按时到达，以满足预定的开始日期。如果一个采购单上的物品比其他采购单上的物品晚到，肯定会导致等待。这是导致 CCI 高库存的关键问题。因此，CCI 希望有一个智能引擎来帮助他们解决此类问题。如果某些 PO 确认晚了，是否有可能在不影响其他工作的情况下推掉一些现有的 PO？该引擎旨在指出可能的推出项目，以减少库存。在此，我们以一个简单的例子来说明该引擎的作用，如图 8 所示。
An online report in JavaScript is designed for CCI’s request, providing the report as shown in Figure 9. 根据 CCI 的要求，设计了 JavaScript 在线报告，提供图 9 所示的报告。
This excel report displays a job ‘WA6246-GY10’ with a list of needed items of shortage. For example, the job was scheduled on 18 September 2015 but different PO with promised date for different items. Some POs are confirmed to be late, so this job has to be push-out. To avoid high inventory, the proposed system will suggest some existing POs for push-out. For CCI’s case, one PO will serve more than one job. Therefore, before pushing out a PO, a check as to whether it will affect other jobs must be done. After full synchronisation, the results are shown in Figure 8. 该 excel 报告显示了一项工作 "WA6246-GY10 "和所需短缺物品清单。例如，该工作计划于 2015 年 9 月 18 日进行，但不同的 PO 有不同的项目承诺日期。有些 PO 已确认逾期，因此该工作必须推出。为避免高库存，建议的系统将建议一些现有的 PO 进行推出。就 CCI 的情况而言，一个 PO 将服务于多项工作。因此，在推出一个 PO 之前，必须检查它是否会影响其他工作。完全同步后的结果如图 8 所示。
Figure 6. Interface of mobile apps for receiving process. 图 6.用于接收流程的移动应用程序界面。
Figure 7. CCI Jobs’ information collected. 图 7.收集到的 CCI 工作信息。
IF
Job WA01 will be later than scheduled start date due to item 2’s late arrival, 由于项目 2 迟到，WA01 工作的开始日期将晚于预定日期、
THEN 然后
Weather we can SAFELY push out item 1 and 3 to new later start date without affecting Job WA02 and WA03. 我们是否可以在不影响工作 WA02 和 WA03 的情况下，安全地将项目 1 和 3 推至新的较晚开始日期。
4.3 Fuzzy inference engine applied for picking process of CCI 4.3 应用于 CCI 挑选过程的模糊推理引擎
In this case study, to find out which type of order picking policy is most suitable for CCI , contextual information is used and analysed. Table 3 shows the input and output variables. The number of orders and the SKU are retrieved from WMS. A timer is used to record the staff picking operation, and the IoT device can sense the number of workers in the warehouse. Integrating the information from the IoT-based WMS, the system can help to analyse which order picking is more appropriate by comparing the priority of strict order picking and batch picking. Table 2 shows the rule of strict order picking and Table 3 shows the rule of batch picking. 在本案例研究中，为了找出哪种订单分拣策略最适合 CCI，使用并分析了背景信息。表 3 显示了输入和输出变量。订单数量和 SKU 均从 WMS 中获取。计时器用于记录员工的拣货操作，物联网设备可以感知仓库中的工人数量。通过整合基于物联网的仓库管理系统中的信息，系统可以通过比较严格订单拣选和批量拣选的优先级，帮助分析哪种订单拣选更合适。表 2 显示了严格订单拣选的规则，表 3 显示了批量拣选的规则。
The fuzzy logic engine was constructed after obtaining the input variable. Using MATLAB, Figures 10 and 11 shows the output variable membership function of the strict order picking and the batch picking respectively. 在获得输入变量后，构建了模糊逻辑引擎。图 10 和图 11 利用 MATLAB 分别显示了严格订单分拣和批量分拣的输出变量成员函数。
Figure 9. Push out suggestions for inventory deduction. 图 9.推送库存扣减建议
Table 2. Rule of the strict order picking. 表 2.严格顺序拣选规则。
Rule 1 规则 1
IF
Number of orders is relative high AND 订单数量相对较多 AND
Then 那么
Priority is relatively low 优先级相对较低
Number of SKU is relative large AND SKU 数量相对较多 AND
Possible time for picking is limited AND Number of staff is small 采摘时间有限，工作人员数量少
Rule 2 规则 2
IF
Number of orders is high AND 订单数量多 AND
Then 那么
Priority is relatively low 优先级相对较低
Number of SKU is large AND SKU 数量多 AND
Possible time for picking is limited AND 可能的采摘时间有限，而且
Number of staff is small 工作人员数量少
Rule 3 规则 3
IF
Number of orders is relative high AND 订单数量相对较多 AND
Then 那么
Priority is low 优先级低
Number of SKU is relative large AND SKU 数量相对较多 AND
Possible time for picking is medium AND 可能的采摘时间为中等和
Number of staff is medium 工作人员数量中等
Rule 4 第 4 条
IF
Number of orders is medium AND 订单数量中等 AND
Then 那么
Priority is medium 优先级为中等
Number of SKU is large AND SKU 数量多 AND
Possible time for picking is medium AND 可能的采摘时间为中等和
Number of staff is small 工作人员数量少
Rule 5 第 5 条
IF
Number of orders is high AND 订单数量多 AND
Then 那么
Priority is low 优先级低
Number of SKU is relatively large AND SKU 数量相对较多 AND
Possible time for picking is limited AND 可能的采摘时间有限，而且
Number of staff is small 工作人员数量少
Rule 6 第 6 条
IF
Number of orders is relatively high AND 订单数量相对较多 AND
Then 那么
Priority is relatively low 优先级相对较低
Number of SKU number is large AND SKU 数量多 AND
Possible time for picking is limited AND 可能的采摘时间有限，而且
Number of staff is small 工作人员数量少
Rule 1 IF Number of orders is relative high AND Then Priority is relatively low
Number of SKU is relative large AND
Possible time for picking is limited AND Number of staff is small
Rule 2 IF Number of orders is high AND Then Priority is relatively low
Number of SKU is large AND
Possible time for picking is limited AND
Number of staff is small
Rule 3 IF Number of orders is relative high AND Then Priority is low
Number of SKU is relative large AND
Possible time for picking is medium AND
Number of staff is medium
Rule 4 IF Number of orders is medium AND Then Priority is medium
Number of SKU is large AND
Possible time for picking is medium AND
Number of staff is small
Rule 5 IF Number of orders is high AND Then Priority is low
Number of SKU is relatively large AND
Possible time for picking is limited AND
Number of staff is small
Rule 6 IF Number of orders is relatively high AND Then Priority is relatively low
Number of SKU number is large AND
Possible time for picking is limited AND
Number of staff is small | Rule 1 | IF | Number of orders is relative high AND | Then | Priority is relatively low |
| :---: | :---: | :---: | :---: | :---: |
| | | Number of SKU is relative large AND | | |
| | | Possible time for picking is limited AND Number of staff is small | | |
| Rule 2 | IF | Number of orders is high AND | Then | Priority is relatively low |
| | | Number of SKU is large AND | | |
| | | Possible time for picking is limited AND | | |
| | | Number of staff is small | | |
| Rule 3 | IF | Number of orders is relative high AND | Then | Priority is low |
| | | Number of SKU is relative large AND | | |
| | | Possible time for picking is medium AND | | |
| | | Number of staff is medium | | |
| Rule 4 | IF | Number of orders is medium AND | Then | Priority is medium |
| | | Number of SKU is large AND | | |
| | | Possible time for picking is medium AND | | |
| | | Number of staff is small | | |
| Rule 5 | IF | Number of orders is high AND | Then | Priority is low |
| | | Number of SKU is relatively large AND | | |
| | | Possible time for picking is limited AND | | |
| | | Number of staff is small | | |
| Rule 6 | IF | Number of orders is relatively high AND | Then | Priority is relatively low |
| | | Number of SKU number is large AND | | |
| | | Possible time for picking is limited AND | | |
| | | Number of staff is small | | |
The priority results of the strict order picking and the batch picking are shown in Table 4 . From the results, it is found that batch picking has a higher value of 0.714 than the strict order picking value of 0.286 . Therefore, the finished goods warehouse of CCI Company should adopt batch picking in the warehouse management in order to increase the efficiency of order picking activities. 严格订单拣选和批量拣选的优先级结果如表 4 所示。从结果可以看出，批量拣货的优先级值为 0.714，高于严格订单拣货的优先级值 0.286。因此，CCI 公司的成品仓库应在仓库管理中采用批量拣选，以提高订单拣选活动的效率。
Number of Orders is relative high AND 订单数量相对较多 AND
Then 那么
Priority is relative high 优先级相对较高
Number of SKU is relative large AND SKU 数量相对较多 AND
Possible Time for picking is limited AND 采摘时间有限，而且
Number of Staff is small 工作人员数量少
Then 那么
Priority is relative high 优先级相对较高
Number of Orders is high AND 订单数量多 AND
Number of SKU is large AND SKU 数量多 AND
Possible Time for picking is limited AND 采摘时间有限，而且
Number of Staff is small 工作人员数量少
Number of Orders is relative high AND 订单数量相对较多 AND
Number of SKU is relative large AND SKU 数量相对较多 AND
Possible Time for picking is medium AND 可能的采摘时间为中等和
Number of Staff is medium 工作人员数量中等
Number of Orders is medium AND 订单数量为中等 AND
Number of SKU is large AND SKU 数量多 AND
Possible Time for picking is medium AND 可能的采摘时间为中等和
Number of Staff is small 工作人员数量少
Number of Orders is high AND 订单数量多 AND
Number of SKU is relatively large AND SKU 数量相对较多 AND
Possible Time for picking is limited AND 采摘时间有限，而且
Number of Staff is small 工作人员数量少
Number of Orders is relatively high AND 订单数量相对较多 AND
Number of SKU number is large AND SKU 数量多 AND
Possible Time for picking is limited AND 采摘时间有限，而且
Number of Orders is relative high AND Then Priority is relative high
Number of SKU is relative large AND
Possible Time for picking is limited AND
Number of Staff is small Then Priority is relative high
Number of Orders is high AND
Number of SKU is large AND
Possible Time for picking is limited AND
Number of Staff is small
Number of Orders is relative high AND
Number of SKU is relative large AND
Possible Time for picking is medium AND
Number of Staff is medium
Number of Orders is medium AND
Number of SKU is large AND
Possible Time for picking is medium AND
Number of Staff is small
Number of Orders is high AND
Number of SKU is relatively large AND
Possible Time for picking is limited AND
Number of Staff is small
Number of Orders is relatively high AND
Number of SKU number is large AND
Possible Time for picking is limited AND | Number of Orders is relative high AND | Then | Priority is relative high |
| :--- | :--- | :--- |
| Number of SKU is relative large AND | | |
| Possible Time for picking is limited AND | | |
| Number of Staff is small | Then | Priority is relative high |
| Number of Orders is high AND | | |
| Number of SKU is large AND | | |
| Possible Time for picking is limited AND | | |
| Number of Staff is small | | |
| Number of Orders is relative high AND | | |
| Number of SKU is relative large AND | | |
| Possible Time for picking is medium AND | | |
| Number of Staff is medium | | |
| Number of Orders is medium AND | | |
| Number of SKU is large AND | | |
| Possible Time for picking is medium AND | | |
| Number of Staff is small | | |
| Number of Orders is high AND | | |
| Number of SKU is relatively large AND | | |
| Possible Time for picking is limited AND | | |
| Number of Staff is small | | |
| Number of Orders is relatively high AND | | |
| Number of SKU number is large AND | | |
| Possible Time for picking is limited AND | | |
Figure 10. Strict order picking. 图 10.严格的订单拣选。
4.4 System evaluation 4.4 系统评估
To evaluate the performance of the IoT-based WMS, a case study is conducted in the manufacturing company. The following data are collected to illustrate the performance of IoT-based WMS in the case company. 为了评估基于物联网的 WMS 系统的性能，我们在一家制造公司进行了案例研究。为说明基于物联网的 WMS 系统在案例公司中的表现，收集了以下数据。
Before IoT-based WMS 基于物联网的 WMS 之前
Average number of order/month 93 平均订单数/月 93
Average number of order finished per month 85 平均每月完成订单数 85
After IoT-based WMS 基于物联网的 WMS 之后
Number of the wrong shipment quad1\quad 1 within 3 months 3 个月内寄错 quad1\quad 1 货物的数量
Time for picking one carton 挑选一箱的时间
Number of carton per order 每个订单的纸箱数量
Manpower required 所需人力
1.86 min 1.86 分钟
After adoption of IoT-based WMS 采用基于物联网的 WMS 系统后
Average number of order/month during the pilot run 84 试运行期间平均每月订单数量 84
Average number of order finished per month during the pilot run 83 试运行期间平均每月完成订单数 83
Number of the wrong shipment during the pilot run 0 试运行期间装运错误的数量 0
Time for picking one carton 54 s 挑选一箱的时间 54 秒
Number of carton per order 140 每个订单的纸箱数量 140
Manpower required 3 所需人力 3
Figure 11. Batch picking. 图 11.批量拣选。
Table 4. Results of order picking and priority. 表 4.订单拣选和优先级的结果。
In the case company, through implement the pilot study, the proposed WMS helps to improve the efficiency of receiving process; enhance the order fulfil performance in the warehouse; enhance the accuracy of the inventory management and improve the warehouse productivity; especially in the picking process. Table 5 illustrates the performance of the warehouse operation. 在案例公司中，通过实施试点研究，建议的 WMS 系统有助于提高收货流程的效率；提高仓库的订单执行绩效；提高库存管理的准确性，并提高仓库的生产率，尤其是在分拣流程中。表 5 说明了仓库运作的绩效。
Table 5. The result of the warehouse operation improvement. 表 5.仓库操作改进结果。
tt : time receive the goods per pallet tt ：每个托盘的收货时间
n:n: number of pallets n:n: 托盘数量
Order fulfilment 订单执行
Order fill rate 订单完成率
96%96 \%
99%99 \%
Order complete: Total order 订单完成：订单总数
Inventory management 库存管理
Order accuracy 订单准确性
99%99 \%
100%100 \%
Order ship to right customer: Total order 将订单发送给正确的客户：订单总数
Warehouse productivity 仓库生产力
Picking order per hour 每小时采摘订单
92%92 \%
100%100 \%
Inventory qty in actual: Inventory qty in system 实际库存数量： 系统库存数量
Total SKU in an order * time per pick one SKU 订单中的 SKU 总数 * 挑选一个 SKU 所需的时间
Category Measurement Before WMS After WMS Equation
Inbound process Receiving 2.54 min 0.96 min (t_(1)+t_(2)+t_(3)+cdots+t_(3))/(n)
t : time receive the goods per pallet
n: number of pallets
Order fulfilment Order fill rate 96% 99% Order complete: Total order
Inventory management Order accuracy 99% 100% Order ship to right customer: Total order
Warehouse productivity Picking order per hour 92% 100% Inventory qty in actual: Inventory qty in system
Total SKU in an order * time per pick one SKU | Category | Measurement | Before WMS | After WMS | Equation |
| :--- | :--- | :---: | :---: | :---: |
| Inbound process | Receiving | 2.54 min | 0.96 min | $\frac{t_{1}+t_{2}+t_{3}+\cdots+t_{3}}{n}$ |
| | | | | $t$ : time receive the goods per pallet |
| | | | $n:$ number of pallets | |
| Order fulfilment | Order fill rate | $96 \%$ | $99 \%$ | Order complete: Total order |
| Inventory management | Order accuracy | $99 \%$ | $100 \%$ | Order ship to right customer: Total order |
| Warehouse productivity | Picking order per hour | $92 \%$ | $100 \%$ | Inventory qty in actual: Inventory qty in system |
| Total SKU in an order * time per pick one SKU | | | | |
4.4.1 Improve the efficiency of receiving process 4.4.1 提高接收过程的效率
Using the IoT-based WMS, it can minimise the time of receiving the goods. As traditionally the finished goods warehouse are using the manual record, therefore there will be time-consuming and the average receiving time is 2.54 min . Using the IoT-based WMS, the receiving process can be streamline as the data can be automatically captured and inputted to the WMS and the average receiving time can be largely reduced to 0.96 min . 使用基于物联网的 WMS 系统，可以最大限度地缩短收货时间。由于传统的成品仓库都是使用人工记录，因此会比较耗时，平均收货时间为 2.54 分钟。使用基于物联网的 WMS 系统，收货流程可以简化，因为数据可以自动采集并输入到 WMS 系统，平均收货时间可以大大缩短到 0.96 分钟。
4.4.2 Enhance the order fulfil performance 4.4.2 提高订单执行绩效
4.4.2.1 Order fill rate. Order fill rate can be highly improved by applying the IoT-based WMS. Using the WMS, the location of the inventories are clearly shown in the system. Therefore, the worker can save the time to complete the order rather than spend an hour to find the inventory locations. From the results, it can show that before implementing the WMS, the order fill rate is 96%96 \% (average 89.5 order finish out of 93 order per month) and after the implement the WMS, the order fill rate is improved to the 99%99 \% (average 83 order finished out of 84 order per month). 4.4.2.1 订单完成率。通过应用基于物联网的 WMS 系统，可大大提高订单完成率。使用 WMS，库存的位置会在系统中清晰显示。因此，工人可以节省完成订单的时间，而不是花费一个小时去寻找库存位置。从结果可以看出，在实施 WMS 之前，订单完成率为 96%96 \% （每月 93 个订单中，平均 89.5 个订单完成），而在实施 WMS 之后，订单完成率提高到 99%99 \% （每月 84 个订单中，平均 83 个订单完成）。
4.4.2.2 Order accuracy. The product misidentification can be minimised using the IoT-based WMS in the finished goods warehouse. As there are lots of SKU number and those SKU number is very similar, it may lead to the worker mistake the item and deliver the wrong goods to the customer. If the goods do not have a shipment, the system promotes up the message to let the worker know that they have picked the wrong items. Eventually, it can reduce the probability of picking the wrong goods due to misidentification. From the results, the order accuracy from 99% (average three month will have one wrong order) can be improved to 100%100 \%. 4.4.2.2 订单准确性。在成品仓库中使用基于物联网的 WMS 系统可最大限度地减少产品识别错误。由于有大量的 SKU 编号，而且这些 SKU 编号非常相似，这可能会导致工人认错货品并将错误的货物交付给客户。如果货物没有出货，系统会提示信息，让工人知道他们拿错了货物。最终，它可以降低因识别错误而拿错货物的概率。从结果来看，订单准确率可从 99%（平均三个月会有一份错误订单）提高到 100%100 \% 。
4.4.3 Inventory accuracy can be improved 4.4.3 库存准确性有待提高
Using the IoT-based WMS, it can reduce the chance of record inaccuracy because of poor handwriting or poor data integrity. The IoT-based technology can help the worker record the information of inventory automatically as the worker just require use the handheld device to scan the goods. As the whole process is not requiring data enter and consequently, the mistake of record manually can be decreased. Before implementing the WMS, the inventory accuracy only 92%92 \%, using the WMS, the inventory accuracy improved to 100%100 \%. Also, though out the Wi-Fi connection, the record of the goods can be updated in a real-time manner. In results, it can help the worker monitor the warehouse in more effective ways. 使用基于物联网的 WMS 系统，可以减少因笔迹不清或数据完整性差而导致记录不准确的几率。基于物联网的技术可以帮助工人自动记录库存信息，因为工人只需使用手持设备扫描货物即可。由于整个过程不需要输入数据，因此可以减少手工记录的错误。在实施 WMS 系统之前，库存准确率仅为 92%92 \% ，使用 WMS 系统后，库存准确率提高到 100%100 \% 。此外，通过 Wi-Fi 连接，货物记录可以实时更新。因此，它可以帮助工人更有效地监控仓库。
4.4.4 Improve the efficiency of order picking 4.4.4 提高订单分拣效率
With implementing the fuzzy logic engine of the order picking process, the process time of the order picking process can be reducing. Using the fuzzy logic engine, the software can generate the most suitable method of the order picking policy and batch order policy was applied in the case study. In the case study, the batch picking can help the worker picking the order in less travel time and hence can improve the efficiency of order picking. Moreover, the redesign floorplan also can help to the same type of group storage in the same area, and the concentrate can reduce the travel distance to the worker. 通过在订单拣选流程中使用模糊逻辑引擎，可以缩短订单拣选流程的时间。利用模糊逻辑引擎，软件可以生成最合适的订单拣选策略方法，案例研究中就采用了批量订单策略。在案例研究中，批量分拣可以帮助工人在更短的时间内分拣订单，从而提高订单分拣的效率。此外，重新设计的平面图也有助于在同一区域内进行同类型的分组存储，集中存储可以减少工人的旅行距离。
In the results, it shows that the time of original one order picking required 4.03 h (130 cartons per order multiply by average 0.031 h pick one carton) and after implementing the WMS, it can minimise to 2.015 h ( 140 cartons per order multiply by average 0.015 h pick one carton). It shows that the warehouse productivity can enhance nearly 50%50 \% using the proposed WMS. 结果表明，原来一个订单的分拣时间需要 4.03 小时（每个订单 130 个纸箱乘以平均 0.031 小时分拣一个纸箱），而实施 WMS 系统后，分拣时间可减少到 2.015 小时（每个订单 140 个纸箱乘以平均 0.015 小时分拣一个纸箱）。这表明，使用建议的仓库管理系统后，仓库生产率几乎可以 50%50 \% 提高。
5. Conclusions 5.结论
The operations of a warehouse are required to change accordingly, due to the complex and high variety of customer orders as well as the demand for real-time information. Therefore, the traditional manual warehouse operation is no longer suitable for manufacturers in the era of Industry 4.0. Thus, an innovative WMS is very important to improve the efficiency and allow customised order fulfilment. 由于客户订单的复杂性和多样性，以及对实时信息的需求，仓库的操作也需要随之改变。因此，传统的人工仓库操作已不再适合工业 4.0 时代的制造商。因此，创新的 WMS 系统对于提高效率和实现定制化订单执行非常重要。
In this paper, the proposed WMS is integrated with the fuzzy logic technique to select the most suitable order picking method, thereby enhancing the efficiency of the order picking process. Through the case study results in this research, it was implied that the WMS could help to provide a better warehouse operation performance regarding both tangible and intangible benefits. For the tangible benefits, it can improve order fulfilment such as order fill rate and order accuracy. Moreover, it can enhance the time of receiving, the inventory accuracy and the warehouse productivity in order picking. For the intangible benefits, it can enhance the packing method, and the inventory can be traceable with RFID. In addition, the morale of the staff can be improved. As this study mainly focuses on practical applications and the order picking operation, the routing and storage policies have not been discussed in detail. However, the space allocation and the reduction of travel distance are very important in enhancing the warehouse performance. Therefore, future work in the fuzzy logic application for batch the zone, sequential zone and wave picking approaches can be further studied. Incorporating artificial intelligence will be one of the future direction to enable smart logistics and information can be further automated so as to streamline the warehouse operation with higher efficiency, performance and less costly in long term and further research on smart robotics should be conducted as it changes the warehouse pick and pack operations from picker-to-goods to goods-to-picker using robots. In general, the adoption of IoT and robotics is the main future research direction to further improve warehouse efficiency. 本文将拟议的仓库管理系统与模糊逻辑技术相结合，以选择最合适的订单拣选方法，从而提高订单拣选流程的效率。本研究的案例研究结果表明，仓库管理系统有助于在有形和无形效益方面提供更好的仓库运营绩效。就有形效益而言，它可以提高订单完成率和订单准确性。此外，它还能提高收货时间、库存准确性和订单拣选方面的仓库生产率。在无形效益方面，它可以改进包装方法，并可通过 RFID 追踪库存。此外，还可以提高员工的士气。由于本研究主要侧重于实际应用和订单分拣操作，因此没有详细讨论路由和存储策略。然而，空间分配和减少移动距离对提高仓库绩效非常重要。因此，未来可以进一步研究批量分区、顺序分区和波浪分拣方法的模糊逻辑应用。融入人工智能将是实现智能物流的未来方向之一，信息可以进一步自动化，从而简化仓库操作，提高效率和性能，降低长期成本。总之，采用物联网和机器人技术是未来进一步提高仓库效率的主要研究方向。
Disclosure statement 披露声明
No potential conflict of interest was reported by the authors. 作者未报告潜在的利益冲突。
Funding 资金筹措
This work was supported by the Collaborative Research [H-ZDAR] and supported by Research Center for China Disability Research, Key Research Base of Humanities and Social Science of Hubei Province [JD18]. 本研究得到湖北省人文社会科学重点研究基地中国残疾人研究中心[JD18]的支持。
References 参考资料
Alyahya, S., Q. Wang, and N. Bennett. 2016. “Application and Integration of an RFID-enabled Warehousing Management System A Feasibility Study.” Journal of Industrial Information Integration 4: 15-25. Alyahya, S., Q. Wang, and N. Bennett.2016."RFID 仓储管理系统的应用与集成可行性研究》。工业信息集成期刊》4: 15-25.
Atmojo, U. D., Z. Salcic, I. Kevin, K. Wang, and H. Park. 2015. “System-level Approach to the Design of Ambient Intelligence Systems Based on Wireless Sensor and Actuator Networks.” Journal of Ambient Intelligence and Humanized Computing 6 (2): 153-169. Atmojo, U. D., Z. Salcic, I. Kevin, K. Wang, and H. Park.2015."基于无线传感器和执行器网络的环境智能系统设计的系统级方法》。环境智能与人性化计算期刊》6 (2)：153-169.
Bajic, E. 2009. “A Service-based Methodology for RFID-smart Objects Interactions in Supply Chain.” International Journal of Multimedia and Ubiquitous Engineering 4 (3): 37-54. Bajic, E. 2009."供应链中 RFID 与智能物体交互的服务方法"。多媒体和泛在工程国际期刊》4 (3)：37-54.
Chow, H. K., K. L. Choy, W. Lee, and K. Lau. 2006. “Design of a RFID Case-based Resource Management System for Warehouse Operations.” Expert Systems with Applications 30 (4): 561-576. Chow, H. K., K. L. Choy, W. Lee, and K. Lau.2006."为仓库运营设计基于 RFID 案例的资源管理系统"。专家系统与应用 30 (4)：561-576.
Culler, D., and J. Long. 2016. “A Prototype Smart Materials Warehouse Application Implemented Using Custom Mobile Robots and Open Source Vision Technology Developed Using EmguCV.” Procedia Manufacturing 5: 1092-1106. Culler, D., and J. Long.2016."利用定制移动机器人和使用 EmguCV 开发的开源视觉技术实现的智能材料仓库应用原型"。Procedia Manufacturing 5: 1092-1106.
Ding, W. 2013. “Study of Smart Warehouse Management System Based on the IOT.” In Intelligence Computation and Evolutionary Computation, edited by Zhenyu Du, 203-207. Berlin: Springer.
