Cross-Docking: Current Research Versus Industry Practice and Industry 4.0 Adoption

Abstract

Abstract Cross-docking is a supply chain distribution and logistics strategy for which less-than-truckload shipments are consolidated into full-truckload shipments. Goods are stored up to a maximum of 24 hours in a cross-docking terminal. In this chapter, we build on the literature review by Ladier and Alpan (2016), who reviewed cross-docking research and conducted interviews with cross-docking managers to find research gaps and provide recommendations for future research. We conduct a systematic literature review, following the framework by Ladier and Alpan (2016), on cross-docking literature from 2015 up to 2020. We focus on papers that consider the intersection of research and industry, e.g., case studies or studies presenting real-world data. We investigate whether the research has changed according to the recommendations of Ladier and Alpan (2016). Additionally, we examine the adoption of Industry 4.0 practices in cross-docking research, e.g., related to features of the physical internet, the Internet of Things and cyber-physical systems in cross-docking methodologies or case studies. We conclude that only small adaptations have been done based on the recommendations of Ladier and Alpan (2016), but we see growing attention for Industry 4.0 concepts in cross-docking, especially for physical internet hubs. Keywords Cross-docking Materials handling Industry 4.0 Physical internet Systematic literature review Supply chain distribution Citation Akkerman, F., Lalla-Ruiz, E., Mes, M. and Spitters, T. (2022), "Cross-Docking: Current Research Versus Industry Practice and Industry 4.0 Adoption", Bondarouk, T. and Olivas-Luján, M.R. (Ed.) Smart Industry – Better Management (Advanced Series in Management, Vol. 28), Emerald Publishing Limited, Bingley, pp. 69-104. https://doi.org/10.1108/S1877-636120220000028007 Publisher: Emerald Publishing Limited Copyright © 2022 Fabian Akkerman, Eduardo Lalla-Ruiz, Martijn Mes and Taco Spitters. Published by Emerald Publishing Limited. This work is published under the Creative Commons Attribution (CC BY 4.0) licence. Anyone may reproduce, distribute, translate and create derivative works of this book (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this licence may be seen at http://creativecommons.org/licences/by/4.0/legalcode. License This work is published under the Creative Commons Attribution (CC BY 4.0) licence. Anyone may reproduce, distribute, translate and create derivative works of this book (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this licence may be seen at http://creativecommons.org/licences/by/4.0/legalcode. Introduction Cross-docking is a supply chain distribution and logistics strategy for minimizing long-term storage of products and parts, maximizing fleet utilization and minimizing trucks dispatches. Over the past decades, the domain has gained attention from a variety of industries, e.g., retail, automotive, perishable goods and pharmaceutics (van Belle, Valckenaers, & Cattrysse, 2012). At cross-docking terminals, cargo is typically stored for less than 24 hours. The process at the terminals consists of unloading, sorting and temporarily storing the goods of inbound trucks from suppliers, after which the goods are moved across the terminal where they are loaded on outbound trucks to be dispatched further down the supply chain. Developments in cross-docking systems have generated competitive advantages. The market of supply chain management is becoming increasingly competitive, making it crucial to optimize logistics costs and throughput of products to stay competitive, as observed in industries such as retail chains, e.g., Walmart (Chen, Fan, & Tang, 2009), and mailing companies, e.g., UPS (Forger, 1995). Logistics-related activities are one of the main cost drivers for many industries (Gue, 2014; Wang, Ranganathan Jagannathan, Zuo, & Murray, 2017). Furthermore, the environmental impact becomes more important; for every voluntary disclosure of an additional 1000 metric carbon emissions, the value of a company deteriorates by $212,000 on average (Matsumura, Prakash, & Vera-Muñoz, 2014). Solution methods for cross-docking that quantify and minimize carbon footprints are under development (Çolak et al., 2020; Nathanail, Terzakis, & Zerzis, 2020). Hence, due to the nature of low inventory levels and higher utilization of trucks, cross-docking has found an increasing amount of attention in the domain of (green) supply chain management (Dadhich, Genovese, Kumar, & Acquaye, 2015). There is a wide range of literature on quantitative methods to optimize various decision levels of cross-docking problems, e.g., truck scheduling (Berghman, Briand, Leus, & Lopez, 2015; Bodnar, de Koster, & Azadeh, 2015; Shakeri, Low, Turner, & Lee, 2012), vehicle routing (Ahmadizar, Zeynivand, & Arkat, 2015; Dondo & Cerdá, 2015), truck-to-door assignment (Guemri, Nduwayo, Todosijević, Hanafi, & Glover, 2019) and the shape of cross-docking terminals (Bartholdi & Gue, 2004). Due to the computational complexity of cross-docking problems, the majority of the proposed optimization models use simplifications. Often, such models do not match industry requirements, and reviews suggest an absence of implementation focus on cross-docking literature (Ladier & Alpan, 2016). To compare what industry practices are commonly used in the literature, Ladier and Alpan (2016) conducted a review study divided into two parts: state-of-the-art literature research on quantitative solution approaches to operational cross-docking problems, and on-site research with interviews at eight cross-docking terminals. The authors constructed a framework for comparing cross-docking research with industry practice. A significant share of the quantitative studies use modelling constraints and performance measures that do not adequately reflect real-world industry practice (Ladier & Alpan, 2016). Ladier and Alpan (2016) recommended connecting future research and industry practice by changing modelling settings and performance indicators. We aim to build on the study performed by Ladier and Alpan (2016). The survey and state-of-the-art review described above was performed in a non-systematic way. The methodology followed in that work involved that selected articles had to be written in English and include well-defined keywords, e.g., cross-dock, cross-docking, transhipment, etc. The authors limited the resulting works to the operational level, resulting in 142 papers up to the year 2015. Since then, numerous new studies have been conducted within the cross-docking domain. Hence, it is interesting to find indications if there is more presence of the practical modelling settings that were recommended for future work. Since 2016, three new literature reviews have been conducted on cross-docking. Buakum and Wisittipanich (2019) conducted a literature review on the period of 2001–2017, exclusively on what type of meta-heuristic solutions are proposed to cross-docking operational problems. The systematic literature review conducted by Ardakani and Fei (2020) extracts information about processes uncertainty for cross-docking planning. Theophilus, Dulebenets, Pasha, Abioye, and Kavoosi (2019) conducted a state-of-the-art review on the timeframe of 2016–2018 regarding the truck scheduling problem. No studies have been conducted on the intersection of research and industry practice. Thus, in this chapter, we conduct a systematic literature review in which we select papers that consider industry practice, e.g., case-based studies or studies presenting real-world data. This allows us to study whether the gap between research and industry practice has narrowed since 2016. Furthermore, we focus on developments in academia and industry with regards to Industry 4.0. The term Industry 4.0 was coined by Kagermann and Wahlster (2011). The term is used for a new revolution in industry, after the first revolution moving from hand labour to machines (1760–1840), the second revolution in faster transport and communication using rail and telegraph (1871–1914) and the third revolution (late twentieth century) in computers and automation. The fourth revolution entails the interconnection of machines, transparency of information, the assistance of machines in human labour and autonomous decision-making of machines (Lu, 2017). Our review examines the degree to which Industry 4.0 has been adopted in academic research and the cross-docking industry. We compare the studies obtained from our review with the state-of-the-art literature before 2015, using the same elements of the classification framework by Ladier and Alpan (2016), i.e., cross-docking settings, business process level and performance indicators. Indications of change or absence of change are used to recommend starting points for future reviews and future implementation-oriented studies. Unlike Ladier and Alpan (2016), we follow a systematic literature review primarily focused on the literature published in the period 2015–2020 and do not conduct interviews with cross-docking managers. Thus, the contributions of this chapter are (1) the summary of new works on cross-docking in the period 2015–2020, (2) the review of changes in literature since 2016, (3) the review of Industry 4.0 adoption in cross-docking research and (4) recommendations for future research in cross-docking and Industry 4.0. The remainder of this chapter is structured as follows. First, cross-docking in the Industry 4.0 era is further explained in Section ‘Cross-Docking in the Industry 4.0 Era’. Next, the methodology for the systematic literature review is explained in Section ‘Methodology’. In Section ‘Results’, the outcomes of the systematic literature review are presented. Finally, we discuss our findings in Section ‘Discussion’, conduct a complementary literature review of Industry 4.0 in cross-docking in Section ‘Extended Search on Industry 4.0 in Cross-Docking’ and draw conclusions in Section ‘Conclusion’. Cross-Docking in the Industry 4.0 Era Cross-docking is typically defined as the process of consolidating less-than-truckload (LTL) shipments with the same destination to full truckloads (FTL), with the additional trait that products are stored up to a maximum of 24 hours (Boysen & Fliedner, 2010). The processes at cross-docking terminals generally constitute unloading of inbound trucks, checking, sorting, temporary storage, transhipment across the terminal and loading into outbound trucks. We give an overview of cross-docking decisions in Section ‘Cross-Docking Research and Industry Practice’. Next, we introduce Industry 4.0 and its potential for cross-docking in Section ‘Industry 4.0 Components in Manufacturing and Logistics’. Cross-Docking Research and Industry Practice Cross-docking allows consolidation of many LTL shipments into fewer FTL trucks, transporting the variety of products or parts in a network of consumers, warehouses and producers. Furthermore, the trait of 24-hour maximum storage time enables a feasible production and transportation strategy for industries with perishable products, e.g., food or pharmaceutical industries. Successfully implementing cross-docking operation strategies into an organization requires changes to business model and operations (Stephan & Boysen, 2011), e.g., finding the operational gains of cross-docking, analyzing network integration suitability and evaluating possible negative effects, e.g., delayed delivery times on customer satisfaction and double-handling costs. Successful implementation of cross-docking into a company's supply chain seeks to eliminate and reduce redundant operations, e.g., storage and movement of products (Enderer, Contardo, & Contreras, 2017). Since the 1980s, several cross-docking studies were done, following the industry trend. Only after 2004, cross-docking started to receive significant attention in the scientific literature (van Belle et al., 2012; Ladier & Alpan, 2016). Over the past three decades, numerous studies have been conducted on the feasibility and benefits of implementing cross-docking for various industries, e.g., the pharmaceutical branch (Ponikierska & Sopniewski, 2017), automotive industry (Witt, 1998), food industries (Vasiljevic, Stepanovic, & Manojlovic, 2013), retail (Benrqya, 2019; Buijs, Danhof, & Wortmann, 2016) and online retail (Cattani, Souza, & Ye, 2014). To benefit from incorporating cross-docking operations in supply chains, it is recommended to adopt a more holistic approach to supply chain management than traditional warehousing (Vogt, 2010). Employing cross-docking operations implies almost eliminating the storage buffer in a distribution network. The local cross-docking operation efficiency is interdependent with distribution activities across the supply chain network. Planning across the entire supply chain is crucial, and the models that include uncertainty are indispensable for coping with disturbances in the supply chain network (Ardakani & Fei, 2020). Introducing cross-docking terminals allows for a reduction in the number of trips for the truck fleet (Buijs, Vis, & Carlo, 2014). Initially, each of the suppliers producing unique products would often directly ship LTL batches of their product to each of the costumers or a long-term storage warehouse. These customers are sometimes located in high-traffic city hubs or other types of urban areas for industries like retail or foods. Last-mile delivery often takes a significant time (Nathanail et al., 2020). By introducing the intermediate stop at a cross-docking terminal, all suppliers deliver shipments less frequently, and the products are consolidated in FTL trucks according to the specific customer demands. Fig. 1 illustrates the difference between classical direct transport (left) and transport using cross-docking (right). Additionally, the decisions for cross-dock location and vehicle routing are indicated. Opens in a new window.Fig. 1. Schematic Representation of a Classical Supply Chain and a Cross-Docking Supply Chain; the Meaning of the Cross-Docking Decision Levels: (a) Cross-Dock Location Selection and (d) Vehicle Routing Decision. Truck scheduling and internal procedures vary per industry. Fig. 2 illustrates a general structure of a cross-docking terminal. The letters in Figs. 1 and 2 represent the decision-making levels for the cross-docking distribution strategy. Opens in a new window.Fig. 2. Schematic Representation of an I-Shaped Cross-Dock, Decision Levels: (b) Design and Terminal Layout, (c) Door Policy and Assignment, (e) Truck Scheduling and (f) Internal Resource Scheduling. Typically, cross-docking problems have been classified into three levels of decision-making: strategic, tactical and operational. We utilize a rephrased definition of the three levels applied to cross-docking as introduced by van Belle et al. (2012). Strategic decisions in this context deal with the location and layout of cross-docking terminals. Location planning is centred around the decisions regarding the structure of the distribution network and the locations of cross-docking terminals. The design and layout of the terminal concern the physical characteristics, i.e., shape and the number of doors. Building shapes are often indicated by a letter, e.g., I, X, L or T. A comparison of the various building design choices of a cross-docking terminal can be found in Bartholdi and Gue (2004). Furthermore, for optimal building selection and cost-to-quality real-estate acquisition, there are models in development that take a company's specific needs and variables into account (Baglio, Perotti, Dallari, & Garagiola, 2019). Tactical decisions concern the design of cross-docking networks. This involves deciding how goods flow through a network that contains more than one cross-docking centre. We refer to the works of Lim, Miao, Rodrigues, and Xu (2005), Chen, Guo, Lim, and Rodrigues (2006), or Ma, Miao, Lim, and Rodrigues (2011) as examples of tactical cross-docking works. Concerning operational planning, the problems at this level relate to vehicle routing, truck scheduling, door assignment and internal resource scheduling. The vehicle routing problem considers combining distribution policies into the network, e.g., direct shipments or milk runs, by flexibly deciding per supplier whether to stop at a cross-docking terminal. We refer to a review of the VRP literature and a case study for multiple cross-docking terminal routing conducted by Nasiri, Rahbari, Werner, and Karimi (2018). Truck scheduling involves decisions regarding the schedule of what trucks are (un)loaded at which dock doors. The door assignment decisions regard the assignment and service of inbound or outbound destinations to specific dock doors. The types of door policies have been classified as follows: exclusive door assignment based on the destination, assignment based on the type of product (e.g., fresh, cooled storage) or based on the type of truck (Stephan & Boysen, 2011). Moreover, classical practices have commonly addressed exclusive door services, where typically inbound trucks can only dock on one side of the terminal and the outbound trucks on the opposite side. A flexible approach of mixed service doors allows both inbound and outbound trucks to be docked at any door. A recent example of research on mixed service doors with promising results has been conducted by Bodnar et al. (2017). Lastly, internal resource scheduling relates to decisions regarding the multiple resource coordination problems in the (un)loading, scanning, transhipment, consolidation and possible value-adding processes inside the terminal. For an example of research on workforce planning integration with internal transport planning for (un)loading, we refer to the work of Tadumadze, Boysen, Emde, and Weidinger (2019). Considering the rich family of cross-docking problems based on the planning level, in this chapter, we limit the collection of cross-docking related works to those addressing internal resource and truck scheduling operations. Regarding the latter, the first classification of truck scheduling is provided by Boysen and Fliedner (2010). Since then, the naming of the various types of truck scheduling in the related literature is found to be inconsistent, where different terms refer to the same type of problem or a general one such as cross-docking scheduling when referring to a specific type of cross-docking scheduling problem (Ladier & Alpan, 2016). Consequently, we use the classification provided by Ladier and Alpan (2016): Truck-to-door assignment: It aims at determining which door each truck is assigned to. Truck-to-door problems are scheduling problems where time is explicitly considered. Truck-to-door sequencing: This type of cross-docking problem considers the order of trucks and their assignment to doors to minimize the average distance the cargo is transported inside the terminal. Truck-to-door sequencing and scheduling: These problems focus on the temporal dimension and do not consider which door each truck is assigned to as long as the maximum number of doors is not exceeded. The distinction between both problems is that sequencing only involves the order in which the trucks are processed, while scheduling explicitly considers the arrival/departure times. For a more in-depth nuance between the types of truck scheduling, the reader is referred to the work of Ladier and Alpan (2016). Moreover, each of the aforementioned problems is dependent on one another in some way, and it is possible to create various syntheses of various levels of decision-making per unique industry. Extensive research has been done on the synchronization of the different decision levels (Buijs et al., 2014; Enderer et al., 2017; Luo, Yang, & Wang, 2019). Industry 4.0 Components in Manufacturing and Logistics In this section, we introduce the general concept of Industry 4.0 and list its different components relevant to cross-docking. After an extensive literature search and classification, Nazarov and Klarin (2020) define Industry 4.0 as ‘the integration of networking capabilities to machines and devices that allows seamless collaboration between the digital and the physical ecosystems for increased efficiencies in the organizational value chains that transform industries and the society for an increased level of productivity and efficiency’. Wagire, Rathore, and Jain (2020) conduct a systematic review and construct a taxonomy of Industry 4.0 research. They found 13 distinct research themes that are clustered in a taxonomy of five principal research areas: Industry 4.0 realization strategies, standards and reference architectures, smart factories, real-time data management and new business models. Ivanov, Tang, Dolgui, Battini, and Das (2021) conducted surveys among researchers to examine the current standing of Industry 4.0 research in Operations Management. They found the following technological aspects of Industry 4.0 in Operations Management: (1) cyber-physical systems/embedded systems, (2) Internet of Things (IoT), (3) 3D printing/additive manufacturing, (4) automated guided vehicles, (5) mobile robots, (6) augmented reality, (7) big data and analytics, (8) artificial intelligence, (9) track-and-trace systems, (10) machine-to-machine communication, (11) cloud services, (12) smart products, (13) blockchain and (14) RFID. The systematic literature reviews in Hofmann and Rüsch (2017) and Garay-Rondero, Martínez-Flores, Smith, Morales, and Aldrette-Malacara (2019) recognize the same Industry 4.0 technology components in (digital) supply chain management and logistics research. Based on the aforementioned surveys and literature review, we synthesize the technological components of Industry 4.0 that are potentially relevant to cross-docking, as summarized in Table 1. Table 1. Synthesis of Industry 4.0 Technological Components for Cross-Docking. Industry 4.0 Technology for Cross-Docking Description Cyber-physical systems/embedded systems Integration of physical processes and computation Internet of things/distributed control Connectivity of machines via the internet AGVs/mobile robots Automated movement or transportation on a pre-defined path Artificial intelligence/big data and analytics The use of modern computing for better analysis and decision-making Track-and-trace systems/RFID/smart sensors Tracking of physical entities using sensors Machine-to-machine communication/networked automation Direct communication between automated machines Source: Based on Hofmann and Rüsch (2017), Garay-Rondero et al. (2019), and Ivanov et al. (2021). Cyber-physical systems integrate computation with physical processes (Lee & Seshia, 2011). It integrates computing, communication, and storage with monitoring and control of entities in the physical world (Sha, Gopalakrishnan, Liu, & Wang, 2008). Inside these cyber-physical systems, a network of machines can be connected using the IoT, which is a concept that entails the connectivity of machines (e.g., 3D printers and AGVs), smart sensors, software (e.g., artificial intelligence algorithms for decision-making) and other embedded systems (Kumar, Tiwari, & Zymbler, 2019). Examples of (potential) adoption of Industry 4.0 in manufacturing and logistics are: (1) the full connectivity of suppliers, (2) logistics and suppliers in a single platform, (3) position-based routing of interconnected vehicles to prevent congestion, (4) fill-level information directly communicated to suppliers using smart sensors, (5) full control of a supply chain using RFID sensors (Hofmann & Rüsch, 2017) and (6) the use of sensors to predict maintenance of manufacturing machines (Lee, Bagheri, & Kao, 2015). In this systematic literature review, we examine the adoption of Industry 4.0 topics (Table 1) in scientific studies about cross-docking. Methodology To find out what cross-docking models have been implemented into practice over the recent five years, we conduct a systematic literature review on studies testing and applying quantitative approaches to cross-docking operational decision levels. This study differs from other recent review studies in the cross-docking domain by explicitly focussing on papers that consider industry practice, e.g., a practical case or incorporating real-world data. The purpose of the study is to extend the comparison framework between industry practices and optimization literature from Ladier and Alpan (2016), with a focus on real-world settings. In addition, we study the degree to which Industry 4.0 concepts have been adopted in research and practice. As indicated in Kitchenham and Charters (2007), we develop a review protocol to guide the identification, selection and extraction process of collected studies. In doing so, we followed the guidelines within the PRISMA method (Moher, Liberati, Tetzlaff, Altman, & The PRISMA Group, 2009) and also considering the COCHRANE handbook (Higgins et al., 2019). For our systematic literature review, five phases are formulated: (1) definition of the review scope and formulation of review questions, (2) determination of search terms, (3) formulation of exclusion and inclusion criteria, (4) analysis and (5) synthesis of findings. In Section ‘Phase 1: Scope and Review Questions’, the review questions are explained. Next, in Section ‘Phase 2: Search Terms’, we elaborate on the search terms and finally, we detail the inclusion and exclusion criteria in Section ‘Phase 3: Inclusion and Exclusion Criteria’. Phase 1: Scope and Review Questions To systematically assess the eligibility of each paper for inclusion, we formulate review questions. First, we address the industry practice with the review question: ‘To what extent does the paper consider industry practice?’ Next, we consider the level of decision-making that is considered (i.e., strategic, tactical or operational) with the question: ‘What planning level is the work at hand addressing?’ The solution method proposed is investigated using the review question: ‘What solution method is proposed?’ Finally, we consider the three questions related to the type of information that needs to be extracted after inclusion: ‘What are the performance indicators utilized?’, ‘What were the cross-docking settings utilized in the solution method?’ and ‘On what data was the solution method tested?’ Phase 2: Search Terms The online databases considered for this study were Scopus and Web of Science. Only reports written in English are eligible for inclusion. The considered timeframe for publications is from 2015 to May 2020, and all source types are eligible for inclusion. After piloting several search queries, we settle on the terms ‘cross-dock*’ OR ‘crossdock*’ OR ‘cross dock*’. These conditions result in 704 studies. At first glance, we find a significant number of publications in the search pool on a biochemical process by the name of cross-docking, a binding mechanism for receptors of proteins and ligands. The selection is then filtered to exclude all the research from the biochemistry domain on the cross-docking process using the search term AND NOT (‘ligand*’ OR ‘protein*’), which reduces the number of studies from 704 to 536. We compare the occurrence of cross-docking keywords of the search pool before and after the search exclusion and find that the number of hits on cross-docking specific keywords (e.g., trucks and logistics) remains unchanged, i.e., no literature is mistakenly excluded. The mentioned search criteria are illustrated in Table 2. After duplication removal, 337 records remain for the screening phase. Table 2. Search Criteria used in this Systematic Literature Review. Search terms for titles, abstracts and keywords (‘cross-dock*’ OR ‘crossdock*’ OR ‘cross dock*’) Filter AND NOT (‘ligand*’ OR ‘protein*’) Timeframe Jan 2015–May 2020 Language English Source type All Document type All Publication status All Phase 3: Inclusion and Exclusion Criteria To enhance the reproducibility and robustness of our review, we formulate and present the protocol for including studies in the final selection (Denyer & Tranfield, 2009; Higgins et al., 2019). We first screen the title, abstract and keywords, and in a second phase we read all full papers. After the first screening phase, 55 studies of 337 remain for in-depth screening. Table 3 shows the inclusion and exclusion criteria used. Table 3. Criteria for Inclusion and Exclusion. Inclusion Criteria Exclusion Criteria Practical case presented Theoretical case without real data Quantitative solution method The cross-docking concept is not the main object of study Targets truck scheduling or internal resource operational decision levels Domain level too wide Published between 2015 and May 2020 Duplicate studies Non-English written papers After the first phase, we read all papers and use the same criteria for exclusion. Specific focus lies on the industry practice that is considered. We exclude a paper when it does not have a case study, on-site study, real-world comparison or real-world data. We also exclude papers that do not use a quantitative solution approach or do not target an operational cross-docking problem. After screening the 55 remaining studies, 25 studies remain in the selection. The outcome of the systematic literature review is summarized in the PRISMA flow diagram in Fig. 3. The final number of selected studies is 25, obtained from the initial 337 records. Opens in a new window.Fig. 3. PRISMA Flow Diagram. Descriptive and thematic features are extracted from the selected records. The studies are thematically classified by answering the review questions. Table