Supply chain management, game-changing technologies, and physical internet: A systematic meta-review of literature

To improve the effectiveness and the sustainability of logistics, the Physical Internet paradigm proposes disruptive solutions. This implies developing an ecosystem of tech-based logistics solutions and supporting methodologies that enable all players in global trade to cooperate. The purpose of this paper is to investigate systematic literature review (SLR) studies to gain detailed insight into how innovative transport technologies, and digitalization initiatives around the Physical Internet development impact supply chains. This paper presents the results of a tertiary study that systematically identified more than twelve thousand articles and selects to review 74 secondary studies on the application of disruptive technologies and the Physical Internet initiative on supply chains from a management perspective. This is complementary to previous reviews, since no one provides a comprehensive and consolidated approach towards the relationship of these three fields. The five-stage systematic review process proposed by Denyer and Tranfield (2009) is followed. As a result, we identify the key activities, knowledge areas and strategies in the supply chain field where the Physical Internet and disruptive technologies interact and are game-changing. Also, we present a conceptual framework that summarises the relationships that exist between relevant disruptive technologies, the physical internet topics, and supply chain key activities. The framework is helpful for researchers and practitioners to find potential technologies to invest in, to assess the potential effects on companies of their implementation, and to support strategic decision-making. The paper concludes with an outlook on future research opportunities from operational, tactical, and strategic perspectives.


I. INTRODUCTION
S INCE the Kyoto Protocol, governments and policymakers around the world have tried to deal with the increasing environmental challenges resulting from the climate crisis, pollution, and deforestation that endanger the sustainability of the planet [1]. However, in recent years, mankind proved that reducing the negative impact of supply chain (SC) operations on society and the environment is a difficult task [2]. Additionally, the United Nations proposed an agenda for 2030 composed of a set of 17 Sustainable Development Goals (SDG), which includes efforts to "make cities and human settlements inclusive, safe, resilient and sustainable" (SDG 11), "ensure sustainable consumption and production patterns" (SDG 12), and "take urgent action to combat climate change and its impacts" (SDG 13). Transport is a key driver in many economic activities that account for about 24% of world CO 2 emissions [3]. Moreover, the accelerating technological changes and the establishment of new players in international trade force local, regional, and global transportation to change. Due to this, there is an increasing interest in developing research around topics such as reverse logistics, closed-loop logistics, green logistics and environmental logistics [4]. Also, there are projects with international cooperation and with multidisciplinary research groups from universities and companies developing disruptive and environment-friendly technologies as well as new policies for Supply Chain Management (SCM) in a global context. On the other hand, as part of the arising of the fourth industrial revolution, the digital revolution is characterized by the convergence of technologies that is blurring the lines between physical and digital worlds. Also, the mix of very different technologies disrupt current and future systems. For example, blockchain has the distributed peer-to-peer nature that can address the weaknesses and vulnerabilities of internet of things in terms of security, reliability and computational effort efficiency [5]. Due to this revolution, the concerns from the supply chain management perspective nowadays include the development of projects that look for developing an interoperable ecosystem of tech-based logistics solutions and for supporting methodologies. It is expected that these solutions enable all players in global trade to co-operate and to react in an agile way to volatile political context, market changes, and climate shifts impacting traditional trades. These projects are looking to speed up the path of the Physical Internet (PI). PI is a game-changing concept in the SCM field that implies the development and combination of technologies to create a collaborative logistic network and face the issues that make the current logistics practices unsustainable [6]. A disruptive technological development is also associated with the concepts of Industry 4.0. Further, according to [7], in Industry 4.0 three aspects can be distinguished: a paradigmatic one, a technological one, and one related to sustainability. Even if these three aspects are also remarkable in the PI, the PI topic is rarely mentioned in Industry 4.0. literature. Several PI literature reviews provide strategies to classify the literature on the PI, present research opportunities, or outline the main PI themes, facilitators and barriers [8]- [11]. Despite this, no review provides a comprehensive and consolidated approach towards the relationship of disruptive technologies, the PI themes and SCM activities. While the common elements of these concepts have been examined, there is still a gap in the literature on understanding the evidence on how SCM activities are positively impacted by applying each of the leading Physical internet components and technologies within a common framework. The latter is the main motivation of this paper. We consider that the understanding of the relations between the physical internet, the supply chain management activities and the disruptive technologies will bring an overview for practitioners to improve decisionmaking during under the transformation promoted by the fourth industrial revolution. Thus, the purpose of this paper is to analyse the evidence and gain concrete insight into how disruptive technologies (e.g. main innovative transport technologies, state-of-the-art tools, technological solutions, and digitalization initiatives), and the physical internet initiative are impacting the supply chain management. The research question (RQ) stated in this paper is: RQ. How are the key activities, knowledge areas and strategies in SCM impacted by the development of PI and disruptive technologies? The RQ encompasses a wide variety of research fields, as well as a multidisciplinary approach. Performing a comprehensive review of the entire (primary) literature on these fields would be a project too ambitious to accomplish. However, we can synthesise research in these areas by reviewing existing systematic literature reviews (SLRs). This allows us the identification of research streams, whose analysis and synthesis help to answer the RQ. In this line of thought, this paper presents the results of a tertiary study, a metareview, of already existing SLRs. An SLR is proposed as the methodology to answer the RQ because it is a replicable, scientific, and transparent approach, which seeks to minimize bias [12]. The originality of this paper is twofold: • From what we can tell, there is no other tertiary study looking to consolidate the relationship of disruptive technologies, the PI themes and SCM activities. • We present a conceptual framework that extends the framework proposed by [10]. This framework summarises the findings and shows which knowledge areas and strategies in the SC are affected by the joint development of disruptive technologies and the PI. This framework is compelling for academics and practitioners in the engineering management field because there is a persisting need to investigate emerging trends as a mitigation strategy against the risk of losing value due to external market disruptions. In fact, the average organization spends about $220,000 of its annual budget to do so [13]. Also, this framework helps readers to gain insights into the PI topic that did not receive much attention in either the Industry 4.0 or SCM literature, so far. The structure of the paper is as follows: The research background of key topics for the development of this study is provided in Section II. Section III presents the methodology for the systematic literature review employed in this study. Based on the results, Section IV shows a conceptual framework illustrating the interaction between PI themes, disruptive technologies and knowledge areas and strategies in SC. Section V shows the research results addressing the research questions, and the implications for academics and practitioners. Finally, Section VI outlines the conclusions of the study, limitations, and future research.

II. RESEARCH BACKGROUND
This section provides a research background about PI, and disruptive technologies. So, the reader has a unified definition of the key concepts used in this paper.

A. PHYSICAL INTERNET
The Physical Internet (PI or π) is a concept conceived at the beginning of the 2010s. This initiative aims to respond to the inefficiencies and non-sustainability of current logistics and SCM practices [6]. The PI is expected to work by organizing transportation of physical goods as data packages are moved on the digital Internet [10]. Ballot [14] describes PI as the application of the Digital Internet principles to logistics networks, while [15] defines it as an SC framework that is based on a network of physical components. Physical components are expected to exchange information to improve the effectiveness, efficiency, and sustainability of SCM operations [15]. Montreuil et al. [16] state that the physical elements that make the PI are π-nodes, π-movers and πcontainers. These elements rest on the principles of universal interconnectivity, encapsulation, standard smart interfaces, standard coordination protocols, logistics web enablers, and open logistics systems [17]. The π-containers are conceived as the unit loads that are manipulated, routed by π-movers, and stored in π-nodes. Due to its novelty, it is still early to perform an analysis of the PI adoption. However, the PI is rapidly gaining relevance in both academic and practitioner circles [10], [11]. This is because there are groups of people and organizations developing and supporting the vision of the PI. The Alliance for Logistics Innovation through Collaboration in Europe (ALICE) is set up to develop a comprehensive strategy for research, innovation and market deployment of logistics and supply chain management innovation in Europe. This group already proposed a roadmap of how the PI will gradually replace the logistics of today [18].

B. DISRUPTIVE TECHNOLOGIES
One of the key enablers of PI or Industry 4.0 is the technologies that allow all players in the value chain to be digitally connected. According to [19], these disruptive technologies not only have the potential to drive fundamental shifts in society, but they also enable simpler and more cost-effective business operations, decision-making and production processes. In the following, the main technologies considered in the further sections of this paper are described.

1) Blockchain
Blockchains are ledgers that record transactions in a trustless environment and are protected by the science of cryptography. The blockchain works as a network of nodes, meaning that each node has the same chain decentralized to its database [20]. One relevant advantage of this technology is that there is no need for a third party to verify the transactions because this verification is decentralized and performed by the nodes (clients) connected to each block. Some additional features of blockchains are that they are immutable, transparent and secure [21].

2) Cloud Computing (CC)
In cloud computing, a pool of configurable computational resources is in virtualized and distributed environments, usually, geographically disperse. The shared resources include data storage, processing power, databases, networks, on-demand environment for developing, virtualization, and software applications. These can be rapidly provisioned and released, with minimal management effort, on an on-demand basis through web-based technologies [22]. Clouds are data and information hubs, providing infrastructure, platform, or software services [7]. Thus, in a potential application in SC and logistics, such platforms receive data from the ubiquitous sensors and analyse and interpret the data for providing users with an easy-to-understand web-based visualization [23]. From this technology, new business models appear, which will be described in section IV-C3.

3) Cyber-Physical Systems (CPS)
A CPS is formed by the integration of computation, networking, and physical processes. It implies signal processing and control of manufacturing processes using computers [23]. So, the CPS uses the information to directly act in the physical world, usually with feedback loops where physical processes affect computations and vice versa [7].

4) Internet of Things (IoT)
The IoT enables the information generation and transmission from objects into a IT system. It means that IoT can provide smartness to objects by the interconnection of sensing and actuating devices, like the Radio Frequency Identification (RFID) technology [7]. Therefore, the IoT can connect individually identified products, machines and people together to provide optimized solutions, through data storage, analysis equipment and decision-making tools [24].

5) Big Data Analytics (BDA)
BDA seeks to produce (from large amounts of data) useful insights or products and services of significant value to executives at different levels, enabling them to develop better decision-making processes [24], [25]. Big data are extensive data sets characterized by the five V's: volume, variety, velocity, veracity, and value. Thus, BDA provides tools to manipulate and process large data sets [7], [24].

6) Artificial Intelligence (AI)
AI purpose is to imitate the human brain and perform decision-making like human beings under various situations and circumstances [26]. Machine learning (Deep Learning and Predictive Analytics are often described as applications of Machine learning), Natural Language Processing and Image Recognition are relevant applications of AI.

7) Autonomous Vehicles
Driverless or fully autonomous vehicles are vehicles that are sufficiently automated that the driver can safely engage in other activities, or that can drive themselves without a human driver. It is a reality on certain transport legs due to the technological revolution in areas such as AI. For road transport, autonomous cars are being developed and tested (e.g. T-pods), while for air-based transport, unmanned aerial vehicles or drones are also being introduced for delivery VOLUME 4, 2016 services [27]. This paper mentions applications and uses of autonomous vehicles with an autonomy level greater or equal to level 2: "On-and off-board decision support" proposed by [28].

III. METHODOLOGY
To address the research question (RQ), the Systematic Literature Review technique is used. We decided to conduct a systematic literature review rather than a narrative review because the latter do not have common frameworks to ensure reproducibility of the studies. Since the objective is to gather and synthesise the literature located, a meta-analysis or a bibliometric review were not considered. Moreover, performing a comprehensive review of the primary literature on the proposed fields would be a project too ambitious to accomplish. So given that we have we found enough SLRs (secondary sources) we perform a tertiary study framed in a Systematic Literature Review scheme. The SLR is defined as a well-established procedure that synthesises research in a systematic, transparent, and reproducible manner. This method provides an auditable trail of the reviewers' decisions, procedures, and conclusions [29]. In this paper, information from different academic documents dedicated to a variety of domains and topics (i.e., disruptive technologies, Industry 4.0, PI) is integrated. The five-stage systematic review process proposed by [12] is followed. These stages are: (i) formulation of the research question, (ii) locating studies, (iii) study selection and evaluation, (iv) analysis and synthesis, and (v) reporting and using the results. Figure 1 describes each of these stages, which are explained below.

A. RESEARCH QUESTION
The general research question (RQ) reflexes the scope, purpose and goals to look for in the SLR. The RQ is: RQ. What are the current and new key activities, knowledge areas and strategies in the SC field where the PI and disruptive technologies are game-changing? Since the purpose of the paper is to illustrate the relationships between three different domains of knowledge (technologies, MTS and IP), this general research question is broken down into the following specific questions. The latter were designed to identify elements and relationships between pairs of concepts and by transitivity find a common framework to answer RQ. RQ1. What are the main disruptive technologies and global initiatives impacting international multimodal freight movement and global transport networks? RQ2. What are the key activities, knowledge areas and strategies in SCM where the PI and disruptive technologies are game-changing? RQ3. What are the main applications of the disruptive technologies in favour of the SCM? RQ4. How are the main applications of emerging disruptive technologies and existing tech solutions in SCM related to PI elements? "intermodal freight transport" OR "physical internet" [#2] "logistics" OR "International cooperation" OR "Multimodality" OR "supply chain" OR "shipping" [#3] "intermodal logistics" OR "synchromodality" OR "silk road" OR ("belt" AND "road initiative") OR "hyperloop" OR "autonomous vehicles" OR "artificial intelligence" OR "Industry 4.0" OR "Transport networks" OR "internet of things" OR "big data" OR "blockchain" [#4] "systematic literature review" OR "systematic review"

C. STUDY SELECTION AND EVALUATION
The choice of documents is done fitting to an explicit selection criterion. According to the type of published document, only journal papers, review papers and book chapters are included. Papers that apply a Systematic Review are included. Only papers published in English are selected, as English is the dominant language in the field of engineering management. A selection based on the relevance in terms of intermodal freight transport is done by analysing their title, abstract and keywords. Thus, documents dedicated to a particular SC (i.e., health care or agriculture), or documents with no thematic or content analysis are discarded. Also, reviews of disruptive technologies which do not present a relevant discussion on the impact or application in the fields of SC or transport are not included. These selection criteria  are applied to ensure the transparent choice and that the selected papers are focused on the topic under investigation. As a result of this selection, the number of documents is reduced to 157. This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication.

E. PRESENTATION OF RESULTS AND CONCLUSIONS
In this section, a descriptive analysis of the reviewed documents is presented. Part of the results presented below are obtained by using the Bibliometrix R-tool [30]. First, the evolution over the time of the short-listed papers published is presented in 2. The oldest selected document date from 2016 and the number of publications is increasing year to year, showing the relevance of the PI, disruptive technologies, and SCM fields. There is a wide-spread contribution of articles over diverse journals. Table 2 shows that 13 journals have published 2 or more papers (cumulating 71.62% of papers). International Journal of Production Research (10 papers), Sustainability (Switzerland) (9 papers) and Supply Chain Management: An International Journal (6 papers) are the most contributing journals. Two of these three journals stand out in the fields of Operations Research, Operations Management and Supply Chain Management for the number of papers published and their cite scores. The other journals have different scopes, being sustainability, information technology or logistics. This emphasises the interdisciplinary character of this study Figure 3 shows the classification by the country related to the affiliation of the corresponding author of each document. The figure summarizes the countries that have published at least two articles, no matter if it is a single country or multiple country publication. The United Kingdom and Austria are the most contributing countries with 7 documents each.
In addition, we read the shortlisted articles and the results are summarized in a conceptual framework presented in Section IV. This framework answers the RQs. Also, a discussion of the findings and the identification of research gaps and future research lines are presented. In addition, in the Appendix A., we present the comprehensive list of documents selected and a resume of the taxonomy.

IV. CONCEPTUAL FRAMEWORK
The results obtained by applying the SLR are summarized through the elaboration of a conceptual framework, which is introduced in this section. This framework provides answers to RQ1-RQ4. The framework shows graphically how is the interaction of knowledge areas and strategies (KAS) in the SC with the PI and the relevant disruptive technologies. This, by positioning each KAS on a grid where the rows are components of the PI, and columns present the disruptive technologies. First, the main PI themes are selected from the PI-Based framework presented by [10]. Thus, KAS in SC can be classified according to its relationship with the following seven PI themes: Modular containers, Vehicle usage utilization, Transit Centres (hubs), Data exchange (seamless, secure, and confidential data exchange), Legal framework, Cooperation Models, and Business models. This classification not only involves the PI components (see Section 2), but also allows the reader to identify the KAS from an operational decision 6 VOLUME 4, 2016 This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and  level (Modular containers) to a strategical decision level (Cooperation Models, and Business models). Based on the taxonomy results of our SLR (see Appendix A.), the most mentioned disruptive technologies are selected, namely Blockchain, Cloud Computing, Cyber-Physical Systems, Internet of Things, Big Data, Artificial Intelligence, Operations Research Methods and Autonomous Vehicles. It is important to mention that Big Data, Artificial Intelligence, and Operations Research Methods are parts within the analytics field. Data analytics techniques are classified into descriptive, predictive and prescriptive analytics [31]- [33]. Descriptive analytics is considered the most basic approach to analytics and is based truly on the principle of classical statistics methods [31]. Predictive analytics involve data mining, machine learning and more advanced statistics to identify patterns in data and convert them into business rules [34]. Prescriptive analytics is used to propose sets of actions based on past events. Prescriptive analytics is mainly associated with optimisation and simulation (named as operations research methods in the conceptual framework), and has special relevance in contexts of uncertainty [34]. Despite the relevance of analytics in the literature, we consider that an exhaustive study of the relationship between the analytics field and SCM is beyond the scope of this paper. Readers interested in frameworks that relate how data analytics shapes SC processes can refer to [35]- [37]. However, since our taxonomy results shows the relevance of Big Data, Artificial Intelligence, and Operations Research methods, we mention them in our framework as technologies in order to show the importance that the reviewed papers give to these specific technologies. In addition, if any other technology has a relevant application, it is mentioned in the framework description (i.e., Augmented Reality, additive manufacturing or the Road and Belt Initiative). Finally, the KAS are identified as a result of the SLR. In the following, a description of the selected KAS and its links with PI themes and disruptive technologies is presented. Figure 4 shows the conceptual framework diagram.

A. DECISION-MAKING TOOLS
In the engineering management context, decision-making is challenging due to the complexity, dynamism, and uncertainty of the environment. For instance, inaccurate demand forecast and a lack of shared information between chain members can lead to the bullwhip effect, unnecessary logistics costs and increased delivery times [38]. Similarly, in transportation, there are significant complexities due to a large number of delivery points, variable demands, or changing travel conditions which are difficult to manage in practice. Also, PI modular containers need to be designed to optimize their use, to efficiently and effectively use cargo handling in transit centres and hubs [10]. Decision-making tools in the SC rely on Operations Research and Management Sciences. Operations Research Methods form a part of the concept of a smart factory and are useful to adopt optimal operations planning which lead to production flexibility and enhanced innovations [23]. These tools are also used in scenarios where optimal decisions need to be taken in the presence of trade-offs between two or more conflicting objectives [39]. Operations Research Methods include mathematical programming, simulation, metaheuristics, queueing theory, multiple-criteria decision making and system dynamics [40]- [43]. In addition, the other disruptive technologies provide insights to improve current Operations Research methods. Also, these generate valuable information for forecasting and planning future actions and behaviours [17]. Concerning data capturing and data processing, the usage of CPSs and IoT can allow companies to get automatic and realtime data from different points throughout the SC [44]. Also, Cloud-Based Systems can provide high storage capacity and high-speed computing, enabling quick and independent access to data from any location [22]. This capability can provide significant support for decision-making and planning and can result in dramatic improvements in real-time problem-solving and cost avoidance [7], [45]. To take advantage of this data, BDA enable companies to have real-time focus and simultaneous analysis of multiple data streams. BDA gives to the organizations the ability to use tools and techniques to analyse SC data in batch-wise, real-time, near-time, or as it flows and extracts meaningful insights for decision-making [31]. Also, BDA methods have the potential to ease communication between automation systems. Thus, the decision-making process can be automated or semi-automated [38], [46]. Given the volume of information, Artificial Intelligence (AI), and particularly Machine Learning (ML), can provide solutions in integrated production management [47], predict the probable backorder products before actual sales take place [48]. Also, it can be used in optimisation, automation, and human support by handling complex problems [49]- [51]. Woschank et al. [52] argue that decision-relevant information would be automatically collected, aggregated, and preanalysed by AI, ML, and Deep Learning technologies. Also, Woschank et al. [52] show that these hybrid decision-making processes outperform purely rational decision-making processes. According to Gupta et al. [53], Operations Research Methods has the ability to identify and provide optimal solutions in a well-defined problem space. Therefore, the challenge lies in providing illustrations that have sufficient expressive power for real-world scenarios and can promise fast and precise solution, and an AI and Operations research methods integration is promising to overcome these challenges. This integration can be applied to different areas of operations, i.e. routing, scheduling, pricing, process forecasting and control, among others. These fields can be supported with the help of AI techniques ranging from case-based reasoning, fuzzy logic, knowledge-based systems, genetic algorithms, and hybrid techniques [53]. Also, techniques for pattern recognition and operations research methods have to be combined more often to adequately incorporate disturbances and changing parameter values into operational, tactical, and strategic decision-making simultaneously [46]. In the reviewed documents we found case studies showing how the implementation of disruptive technologies can improve the Operations Research Methods and, therefore, decision-making task in: waste management, water supply networks, power plants [54], food SC [55], resource allocation [56], supply chain network design [39], scheduling, purchasing, procurement, physical distribution [46], traffic information analysis in real-time [38], [54], [57], route planning and refuelling, pricing negotiations, maintenance times and packages flow optimisation [54], [56], [58], [59]; maritime logistics chain optimisation, fleet optimisation [60], truck arrival time prediction, blood SC [31], forecasting demand [43], [57], [61], product development [34], virtual design and simulation of processes [19], [60], [62]; predictive maintenance, human-expert-driven decision-making and AIdriven decision-making approaches for smart manufacturing processes based on AI and ML [52]. In general, integrating external data lead to real-time enabled response modelling and would provide the necessary data for more holistic and realistic models. In this way, Smart goods, vehicles, and infrastructure are keys enablers. In the proposed framework, decision-making tools are embedded in fleet management, automated guided tasks, warehouse location/management and inventory management. Also, these play a key role in real-time information processing and cooperation strategies.

1) Smart goods, vehicles and infrastructure
Data is scattered, and diverse sources can provide it. Usually, the primary data sources are Enterprise Information Systems, Enterprise Resource Planning systems and interorganisational systems. Besides, IoT devices, sensors, and RFID devices can collect data and feeding it to information systems, which in turn generate a huge volume of unstructured data [31], [34], [63]. This embedded cyber-physical intelligence and IoT technologies can be applied in SCM to enhance visibility through enabling monitoring and tracking capabilities. In particular, the combined use of RFID technology, BDA, Blockchain and cloud-based systems can empower real-time traceability of goods and interoperability of resources, leading to improved planning and controlling throughout the SC [38], [44], [63]- [65]. Also, IoT-assisted data-driven digital twins can be integrated with reinforced learning platforms to assist truck route optimisation, automated warehouse storage and retrieval systems, decentralized scheduling, risk management, intelligent transportation system, and other micro and macro reverse logistics procedures [49], [66]. These operations research methods can be used in real time via a cloud computing platform [17], [22], [23]. Real-time traceability has more benefits than just improving the visibility of goods at the data level [7], [67], [41], [64], [65]. Better visibility and information accessibility can reveal opportunities to minimize lead times in warehouse and transportation tasks, to improving manufacturing flexibility, product quality, energy efficiency, and to improve equipment service and communications with customers, leading to a reduction in costs and increases in efficiencies [68]. Recent literature reveals a tendency towards the design and activeness of the PI modular containers, with an emphasis on data sharing connectivity structures and visibility improvements. Using smart tags, including RFID and GPS technolo- This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2022.3181154 This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ gies, PI modular containers collect and store logistics and SC information. In addition, modularisation, multifunctional load units and advanced encapsulation processes are studied as alternatives to increase the fill rate [17]. The activeness of PI modular containers results in better grouping strategies and real-time optimisation to decrease empty containers via cloud computing. However, all these benefits entail challenges related to information security, new regulation requirements, complex network implementation, high implementation costs, hardware and software issues, standardisation and data handling [7], [69]. Also, others risks as low data quality, distrust and trust management, economic risk or complex network coordination can emerge [69].

2) Fleet management
Performance of transport logistics can be increased by applying AI-based methods in combination with state-of-theart approaches based on the usage of information technology, data communication technology, electronic sensing technology, global positioning technology, geographical information system technology, real-time road status, navigation, computer processing technology, and system engineering technology [38], [52]. Also, auto-control and intelligent regulation of connected vehicles effectively reduce time spent on commuting and energy consumption [63]. Studies on BD and IoT in the context of logistics show improvement opportunities like fatigue management for driver safety, predictive maintenance, on-time product delivery, transport route optimisation, truck sharing and transport planning [7], [11], [38], [70]. For last-mile deliveries, a combination of trucks and drones has been explored in recent research efforts. Drones are found to be highly cost-effective and environmentally sustainable if optimized route planning strategies. In this regard, there are recent applications of Operations Research Models for drones' routes optimisation (to and from trucks) with the objective of time and cost savings [71]. Concerning fleet management on maritime transport, financial fleet management can be improved by using artificial neural networks to predict the freight market of tankers, and AI and Big Data applications in navigational avoid vessel collisions by integration with the vessel traffic services. Other contributions are on early detection of vessel delay based on a data-driven method using a large amount of vessel tracking data from real-time satellite-automatic identification systems [56]. In general, vehicle usage optimisation research proposes the idea of using shared, fully loaded, energy-efficient PI vehicles with relays, which reduces transportation costs and carbon footprints. The vehicle usage optimisation theme is part of the PI literature, and contributions to this theme focus on the design and assessment of PI-enabled solutions for the optimisation of vehicle usage [10].

3) Automated guided tasks
Autonomous vehicles in conjunction with automatization and robotics enable performance, efficiency, efficacy, response time increases, and waiting times reduction throughout the SC [38], [72]. Autonomous vehicles extended with other technologies like IoT or CC can be used as cyber-physical systems (Automated guided vehicles) [7]. In addition, in the literature, there is special attention to applications on maritime logistic operations in the shipping industry, from basic operations to anchoring systems [56]. Bȃlan [40] present some contextual factors that lead to a disruptive impact on maritime transport. In the case of autonomous or unmanned ships/vessels, the author points out factors like the challenges for autonomous ship/vessel handlers in supporting situational awareness, as well as the harmony between ship/vessel, personnel and environment. Another challenge to consider is the skill loss by the ship/vessel handlers/crew members, which may follow the implementation of unmanned technology [72]. However, there are some cost changes in the case of autonomous ships/vessels compared to the conventional ships/vessels: more effective and efficient planning and dispatching strategies at platoon hubs will determine financial gains [71]; cost savings in terms of crew wages and crew related costs because of the absence of crew on board; cost increase related to personnel costs and operating costs of the shore control centre and to the cost of maintenance crews that improved design of the ships/vessels with direct influence on fuel consumption and greenhouse gas emissions; and the anticipated overall design of unmanned ships that will be significantly different in terms of hull design and propulsion arrangement [40].

4) Warehouse location and management
Among the latest technological trends, those related to the use of analytical formulation and optimisation algorithms for the management of warehouses are viewed as crucial in logistics systems [41], [73]. Demand estimation and inventory optimisation can be carried out together in one single step by applying machine learning algorithms combining traditional operations research algorithms [57]. On the other hand, PI transit centres encourages the move from non-standardized proprietary transit centres to modern and open locations for efficient and effective cargo handling. Using coordination algorithms for matching demand and supply, the mission of PI transit centres is to efficiently and sustainably transfer PItrailers from one truck to another [10]. In warehouse location/management, large infrastructure projects as the Belt and Road Initiative (BRI) or the Rail Baltica railway line have significant impacts [74], [75]. BRI project aims to revive the historic Silk Road trade route connecting Europe to East Asia via Central Asia. The BRI contributes to developing connectivity, supply chains, and international logistics along the Belt and Road because intermodal transportation and global logistics system go hand in hand [74]. Due to the magnitude of the project, BRI will have positive aspects on factors like labour cost, availability and stability of infrastructure, and market, supplier and knowledge access [76]. Also, additive manufacturing allows the concentration of the production of various lowvolume, high customisation, high urgency parts [77]. So, this type of infrastructure projects and technologies will affect the characteristics and attractiveness of the different regions around the world, which is a determinant factor when making location decisions in a fully collaborative network. In these large infrastructure projects, decision support systems that embed operations research method can provide alternatives for the design and construction of the transport lines [75]. Concerning warehouse management, data transmitted from RFID integrated with cloud-based warehouse management systems can create remote positioning methods and indoor/outdoor GPS-based systems [44]. Also, CPS enable the automated warehousing systems that they are suited for application in many types of transport and material handling systems [66]. Also, CC has a positive effect on warehouse integration, as it integrates both the information flow and the physical flow by providing flexibility and agility, and facilitating resource sharing among participants throughout the SC life cycle [22]. The industrial deployment of IoT infrastructure enhances the order fulfilment platforms or collaborative warehouse platforms. Warehouse visibility, traceability, and transparency can be improved to facilitate competitiveness in a dynamic environment by utilising these ideal platforms for decentralised warehouse management [63]. Thus, implementing disruptive technologies on warehousing management boosts the SC performance thanks to an optimized physical asset, service sharing, and a reduction in warehousing cost [7], [38].

5) Inventory management
The integration between IoT, BDA and Cloud-based systems allow stock reduction, errors decrease and inventory management improvement [7]. CC can contribute to supplier integration, apart from resource sharing, it enables the exchange of skills, know-how and production data. CC affords the company and suppliers real-time access to production and logistics data, and greater SC visibility and flexibility [22]. Improving forecast models, including anti-collision algorithms for RFID detection and automatic object localization reduce cost [7], [39]. Also, a better forecast of product demand not only diminish the uncertainty (thus the safety stock) but also allows planning preventive maintenance and reduced downtime of physical resources [78].

B. REAL-TIME INFORMATION PROCESSING
Under the PI perspectives, the seamless, secure and confidential data exchange research defines a set of open, shared and secure protocols for data exchange in PI-enabled open logistics networks. The latter implies to develop multiple data models, including canonical data and enterprise application integrations, to define a common set of data and information for information exchange and interoperability between par-ticipants in open logistics network [10]. This forces companies and data-sharing applications to improve, the capability to handle a large amount of real-time data and the end-to-end SC visibility. This can be done by including disruptive technologies. The adoption of Internetbased inter-organisational systems and Internet-based electronic data interchange has enhanced communication, coordination, and collaboration across organisational boundaries [31]. In this way, developing Big Data Analytics capability became mandatory for building competitive and sustainable SC [79]. This implies important challenges on the value of the information and how all this data should be shared and processed. Concerning information sharing, Big Data technologies and CC have modernised traditional mechanisms between SC members [31], [56], [68]. Three types of changes can be observed: horizontal integration across the value chain, endto-end engineering and vertical integration of networked manufacturing systems within the organization [19], [80]. In this sense, information transparency and proper communication between SC echelons, and the development of a single set of common standards to support collaboration are relevant enablers to real-time information processing [23]. Also, cloud-based collaboration increases technological proximity by integrating multiple sources in a system, so customers can use applications and services in the cloud, reducing the risk of out-of-date and incorrect information [22]. Besides, CC technologies enable storing a large volume of data and real-time monitoring, with information accessible to all SC members anywhere at any time [55], [78]. In terms of the benefits of instantaneous data processing, realtime information transmission between buyers and sellers facilitates the integration of physical flows, which affords greater agility and flexibility to respond to fluctuations in demand, as well as reduction in planning time, enhanced customer satisfaction and optimisation of decisions for reverse logistics [7], [31], [38], [44], [78], [81], [82]. Also, customer satisfaction increases can be achieved by constant communication, information flowing from the service provider and price competitiveness [83]. In addition, BDA and Blockchain encourages ethical behaviours among SC tiers because of traceability, transparency and moderate uncertainty by guaranteeing better quality and accuracy of the activities, which may increase commitment to sustainable practices and also the reputation of industries [55], [79], [84], [85]. Some applications mentioned in the literature are: to support navigation and scheduling tasks and to improve order assignment to containers [7]; IoT devices paired on blockchain networks to fight food fraud, counterfeit medicine and luxury jewellery like blood diamonds [21], [86], [87]; Realtime monitoring used to protect the environment because optimizing maritime traffic reduces the risk of accidents and minimizes environmental pollution [60]; and Blockchainenabled real-time information tracking has applications as smart contacts, or document sharing and version control [86]. Despite all the benefits, the amount of shared information can cause a bottleneck produced by aspects like the low bandwidth and high latency of 3G/4G [88]. Thus, [88] argue that the 5G is expected to resolve those issues and to connect the SC partners for sharing those data in real-time. According to [89], 5G-enabled IoT increases the bandwidth capacity for the secure transmission of goods-related data.

1) Risk management
In the risk management field, there are two types of contributions. The first assessing how disruptive technologies help to improve the current risk management activities, while the second type focuses on the emerging risks of implementing disruptive technologies in the operations. BDA gives to organizations the ability to distinguish between risks that must be avoided and risks that must be taken. This is possible by identifying trends and events through monitoring publicly available news or social media channels associated with suppliers or specific sourcing markets [33]. Thus, organizations can continuously obtain updated information on suppliers and sourcing markets and quickly respond to changes or supply risks, even with contingency plans. Using blockchain technology for tracking cargo allows faster processing of insurance claims in cases where cargo has been lost or damaged, because tracking data on blockchains are trustworthy and traceable to the origin of loss [21], [90]. Thus, insurance companies can process the causes of the incident, the carrier involved, the type of cargo and the validity of the claims faster and easier. Also, ML can be used to model disruptive events and their impact on the supply chain to identify potential risks in a timely manner [91]. In maritime fleet risk management, the assessment of ship risks can be improved by: detection of anomalies in marine operations from data gathered on vessel movement; investigation of cargo loss in logistics systems employing datadriven analytics; and identification of how CC may become an enabler of dynamic and synchro modal container consolidation [40], [56]. As mentioned before, real-time data exchange and BDA methods combined with AI provides responsiveness, agility and flexibility to SC. However, it is important to guarantee transparency [67], [68], [92]. According to [83], data sharing in an intermodal network is complex, as the information systems used might be incompatible or there could be issues with speed, management capacity, the volume of information shared and the fear of revealing explicit information about the firms involved. For this reason, the influence of technological solution service providers could play a strategic role by promoting greater information and communication technologies adoption by network actors. Thus, transparency induces proactive behaviour in the SC members by enabling them to identify and respond to various changes and potential disruptions [67]. In this sense, the use of IoT technologies contributes to increasing reliability thanks to the elimination of information processing errors [38]. Also, from a systemic perspective, blockchain may leverage its potential through a large-scale collaboration of stakeholders as SC friction mainly stems from dispersed disconnection among parties [93]. Blockchain can be used to solve the issues related to double marginalization and information symmetry in the SC, by providing formal guarantees to the parties [94], [95]. Using IoT technology for real-time communications usually exposes smart devices to a range of security risks. So, the cybersecurity of each SC member is an issue that can decrease the level of trust [21]. Also, an issue with having a centralised cloud system for all IoT devices within an SC is its susceptibility to cyber-attacks that can make SC services unavailable until the cyber issue is eliminated [54]. From a data safety and integrity point of view, and due to the high cost of maintaining centralised IoT systems in SC and the security concerns surrounding IoT devices, one alternative is the application of blockchain networks [20], [42], [93]. Blockchain, as a distributed shared ledger technology, may help increase traceability and extend SC visibility by its consensus mechanism [93]. In this case, the transactions between IoT devices are protected by cryptography and are verified to ensure the originator of the message is not a malware or external intermediary [21], [42], [89]. Thus, decentralisation is a unique data security mechanism of the Blockchain [21], [54], [87], [94]. Regarding traceability, Blockchain sees large-scale deployment because a block could be created for each transaction following the product's digital footprint, from manufacturing to distribution and sale. Thus, every transaction along a Blockchain Supply Chain is fully auditable [87], [96], [97]. Even if Blockchain in the long term would yield to cost reduction, there are immediate implementation costs in the early adoption as those associate to incompatible blockchain models or upfront high costs for "mining" [94]. So, returns can occur at different times depending on the particular objectives of each application.

2) Cooperative inventory management
Integration and interoperability allow SC members to work closely [67]. In vertical integration, functions inside companies work more effectively with technologies such as Enterprise Resource Planning being key to such integration. While horizontal integration can be achieved across the SC with Cloud Technologies, IoT and Digitalisation. To achieve a high level of cooperative inventory management, the vendor managed inventory system uses, in conjunction with IoT, rapid and coordinated real-time inventory management across the SC and so permits a minimum yet nonetheless flexible inventory level that can address customer demand fluctuations [38]. According to [33], the use of BDA in vendor managed inventory systems collection, processing, and reporting on inventory data can inform decisions related to inventory performance improvement. Also, BDA helps to obtain a holistic view of inventory levels across the SC, while considering the impact of inventories at any given level or echelon. In consequence, it can help in decisions related to safety stock optimisation. The utilisation of the IoT in vendor managed inventory activities supports transparency, agility and efficiency by real-time and bidirectional information flows [7].

3) Transportation cooperation
A freight transportation network includes multiple players (and multiple modes of transport) that do not necessarily trust each other and, in most cases, do not have a standardised method of sharing the transport data. In this case, using a customised blockchain that connects these parties would ensure sharing the required transaction and shipment data, which are both secure and reliable given the inherent features of the Blockchain [21]. Likewise, CPS, through enabling a high level of integration and information exchange, can enable a better understanding of the requirements of different parties and can enhance collaboration and cooperation between them [44]. According to the PI paradigm, it is expected that transport is organised and optimised in a decentralised way. In other words, for a given request, its best route from the origin to the destination will be updated every time it arrives at a PI-hub according to real-time, local information. To manage such decentralised systems, transport protocols and collaborative protocols will play a relevant role in guaranty the level of service and global optimality of the network [70]. In this vein, large infrastructure projects as the BRI must be taken into consideration. This connectivity comprises infrastructural, trade, and financial aspects [76]. From an infrastructural perspective, the BRI could be expected to contribute to establishing more resilient SCs, such as by increasing the quality and dependability of the logistics infrastructure [98]. BRI may enable other strategies for enhancing resilience, such as increasing visibility and enabling greater SC collaboration through greater connectivity. The latter is possible with the right implementation of the disruptive technologies presented in the framework [76]. A couple of studies on applications of transport cooperation are found by [70]. One, which investigates how information communication technology can help carriers dynamically show best transport plans by sharing real-time information. And the other, which investigates how shared information employed between collaborating shippers can help reduce CO2 emissions from freight transport in the grocery retail industry in the UK. Also, there are models where product distribution and delivery are underpinned by CC. These models show that it is possible to offer increasingly better delivery results as the number of participants in a collaboration environment grows [38].Finally, under the PI perspective, the cooperation models research attempts to redefine the existing practices for revenue sharing among different stakeholders in the new PI-enabled business models, such as PI hub holders and PI movers [10].

4) Smart contracts
In PI, the legal framework research aims to synchronize the incompatible legal environments associated with different countries to provide legal security and seamless international transport [10]. In this vein, the smart contracts become an important tool. Smart contracts are programmable protocols that allow the execution of contract terms and agreements [87], [93]. This operation does not rely on an intermediary. Therefore, it not only speeds up the transaction but also promotes costs reduction and improves trust, since, within the network, all participants (nodes or actors) have a copy of the ledger [99]. The goal is to achieve effective and efficient flows of products and services, information, money, and decisions, to provide maximum value to the stakeholders [20], [87]. Smart contracts in transport and logistics, coupled with tracking tools, can facilitate payments to suppliers or 3PLs once they fulfil their tasks such as delivering goods to a warehouse or a port in a predefined specification [21]. Thus, the tracking device at a buyer's warehouse is connected to the Blockchain and once it receives the cargo from the supplier, it checks the quantity and the quality. This verification can be handled by using smart contracts running on 5G to track a shipment [89]. If all the stipulated conditions are met, the smart contract can automatically release the payment to the supplier [21]. Also, smart contracts can be used across cold chains to ensure that desirable conditions are maintained during manufacturing and transportation as well as issuing warning signs in case the sensors report any abnormalities across the chain [21], [89]. By utilizing a smart contract feature to execute digital signatures, the validation processes are speeded, because BCT provides trusted data from a single source. Enterprises may benefit from the reduced processing of paper-based documents as well as saving the considerable costs involved in tracking and obtaining proof of information authenticity [20], [73], [86], [93], [100]. Besides, smart contracts have the potential to solve or alleviate the problem of information asymmetry [95].

C. NEW BUSINESS MODELS
According to Treiblmaier et al. [10], up to 2019, business models were the most mentioned theme, on the physical internet literature. With the increase in the number of publications and the joint review of reviews of technologies and Physical Internet, we found more focused contributions. So, in the following we present the insights around new business models clustered in four main topics.

1) Advanced automation
IoT allows the developing of self-monitoring capabilities where machines and devices can monitor and communicate their real-time performance [44]. This feature combined with intelligent devices, systems and production processes enable advanced automation throughout the entire SC that can lead to improved productivity efficiency and quality control [38]. This automation of processes results in less workforce requirement and process efficiency improvement [78]. Also, some articles show opportunities for using CPS and IoT regarding decentralisation and efficiency. Nonetheless, it is 12 VOLUME 4, 2016 This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2022.3181154 This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ discussed that while higher degrees of autonomous control support goal achievement in logistics, although too much decentralisation could lead to obstructive chaotic systems [7].

2) Cloud manufacturing
Because of the increasing use of sensors on physical products, there is a vast potential for the development of sensorbased applications. For these reason, cloud manufacturing and IoT are interlinked. These applications may be used for preventive maintenance, to avoid stock-outs through monitoring inventory levels, for better capacity planning, and to assess the usage and functionality of products [68]. Other cloud manufacturing approaches, in productions logistics, are reported as part of [7] review. They show a case study where thanks to the synchronisation of an IoT-enabled production environment and a cloud supported resource management, it is possible to find improvements in delivery rates and inventory levels.

3) Service models
Even if most of the applications of CC that we found are related to data ingestion. Capabilities as storage, processing power, databases, networks, and software applications are key enablers of BDA capabilities. According to [22], [31], three different types of service models can be distinguished in CC. Infrastructure as a Service involves sharing data or IT infrastructure that can be used as a service. Platform as a Service entails providing a complete platform for application development and deployment on demand. Software as a Service where cloud providers host and manage the software application and underlying infrastructure, and handle any maintenance, like software upgrades and security patching. These services can be further deployed in four different ways: private cloud, community cloud, public cloud, and hybrid cloud. Recent studies state that cloud-computing services such as IaaS can be deployed on a community level to facilitate data sharing between SC partners within the network. Also, these studies argue that several business models based on CC such as Analytics as a service, Big Data as a service and Knowledge and information as a service will enhance BDA [31].

4) PI-enabling firms
Several publications outline the PI as a key driver of business model innovation [11]. Thus, the actors in the logistic landscape can be divided into PI-enabling firms and PI-enabled firms. For example, virtualisation of supply chains enables to decouple physical flow from coordination and planning. It thus becomes a challenge to optimise flow in virtual supply networks that dynamically change their configuration depending on the state of the physical supply chain system [64]. In the same vein, the concept of digital twin shopfloor, based on the convergence of the physical and virtual worlds of the shop-floor, requires adapting the optimisation and forecasting actions [61]. New business models might be focused on auctioning, transit centre management, or less-than-truckload dynamic pricing in the PI. In general, the authors outline the need for transparency in PI business models to avoid principal-agent conflicts [11].

V. DISCUSSION
Based on the findings presented in the previous section, the research questions posed at the beginning of this review are answered. Below, we present the research results, as well as the implications for academics and practitioners offered from our tertiary study.

RQ1.
What are the main disruptive technologies and global initiatives impacting international multimodal freight movement and global transport networks? The main disruptive technologies and global initiatives identified in our review were Blockchain, Cloud Computing, Cyber Physical Systems, Internet of Things, Big Data, Artificial Intelligence and Autonomous vehicles. However, other technologies and initiatives as The belt and road initiative, the augmented reality and the additive manufacturing were identified in a smaller proportion compared to those selected in the conceptual framework. The comprehensive list of documents selected and a resume of the taxonomy used is presented in Appendix A. As mentioned before, among these technologies, there are elements of the analytics field (Big Data and Artificial Intelligence). This is the reason why studying analytics and its relationship to SCM yields the attention of researches [35]- [37].
RQ2. What are the key activities, knowledge areas and strategies in SCM where the PI and disruptive technologies are game-changing? In the conceptual framework presented in Figure 4, we consolidate and summarise the findings in three clusters with subclusters. First, the decision-making tools which impacts the management of the smart goods, vehicles and infrastructure, the fleet, the automated guided task, the warehouse and the inventory. The second cluster, where the real-time information processing and sharing impacts the risk management, the cooperative inventory management, the transportation cooperation and the smart contracts. Finally, the third cluster on the new business models that are going to create or modify the roles of the SC players. There we identify subclusters of new business models in advanced automation, cloud manufacturing, as a service models and PI enabling firms.
RQ3. What are the main applications of the disruptive technologies in favour of the SCM? The description of the conceptual framework, presented in Sections IV-A -IV-C, presents in detail the application of the selected technologies in each of the KAS subclusters identified. Results in Section IV-A show that the integra-VOLUME 4, 2016 13 This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2022 This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ tion of disruptive technologies led to enhance the global performance of an SC. Each technology provides a special characteristic not also to SCM but also to PI themes. Thus, IoT offers the possibility of generating a huge amount of data [44]. BDA brings the tools to analyse and extract valuable information from the collected data [31]. CC relates to the infrastructure where the data can be managed and gives the desirable interconnectivity [22]. Cyber-physical systems enhance productivity by enabling an elevated level of integration and information sharing [66]. Blockchain is seen as a key technology to mitigate security and transparency issues [92]. However, several applications need some minimum connectivity requirements that are expected to be solved by the implementation of 5G connectivity worldwide [88]. Further, a fully connected network would enable not only to minimise costs and time, and maximise users' profits, but also to increase safety and to fight illegal activities (see Section IV-B) [21], [86], [87]. Nevertheless, access to information, top-notch technology and infrastructure investment capabilities are essential to boost the implementation of the PI initiative worldwide. The changing structure of supply networks strengthens the need for collaborations and therefore, the urgency for efficient communication and data exchange [10], [44], so there is expected to be a significant development of new business models that will enable these transformations (see Section IV-C) [10].
RQ4. How are the main applications of emerging disruptive technologies and existing tech solutions in SCM related to PI elements? In figure 4, we draw the relations with the PI themes in such a way that the reader can identify how the elements are impacted from by KAS in an operational decision level, in a tactical decision level, and in a strategical decision level. In the operation decision level, the modular containers can take profit the of the integration of disruptive technologies in the decision-making tools, and of the management of smart goods, vehicles and infrastructure. Vehicle usage utilization is highly related with the fleet management and the automated guided task, which includes the management of autonomous vehicles. Likewise, the transit centres are related to the automated guided task, but also to the warehouse location and management and the inventory management. In the tactical decision level, data exchange is enabled by the technological development around the real-time information processing and can be supported by the risk management field. The cooperation models in PI are impacted by the risk management, the cooperative inventory management, the transportation cooperation and the implementation of smart contracts. In this case, disruptive technologies helps to overcome scalability, security and reliability issues. And in the strategical decision level, the legal framework theme is related to the smart contracts implementation, which allows the players to reduce the number of intermediaries and therefore to speed up the cooperation and the transactions between players. Finally, the development of new business models around the PI are aligned to the development of advanced automation, cloud manufacturing, service models and PI dedicated firms.

B. IMPLICATIONS FOR ACADEMICS
For researchers, this study sheds light on the PI and provides new insights into the relationship between disruptive technologies, PI components and SCM. The conceptualisation of these relations helps academic researchers to embark on new empirical research in this domain. The proposed framework offers intuitions into possibilities to enhance existing logistics systems and improve technological solutions. Moreover, this framework exposes the importance of multidisciplinary research groups working on the design of the future of freight transport.

C. IMPLICATIONS FOR PRACTITIONERS
For practitioners, this paper shows advantages, initiatives and risks related to implementing certain technologies on the transportation tasks and how these contribute to the PI, as well as how these allow for improvements in the performance of SCs, which is an essential field in the engineering management. Also, as countries/companies invest in the development of the PI, it is relevant for managers to have exposure to such concepts and to be updated on the development of disruptive technologies for the future of transportation worldwide. Likewise, the proposed framework is useful to assess the potential effects on companies and to support strategic decision-making.

VI. CONCLUSIONS AND FUTURE RESEARCH
This paper studies literature reviews around disruptive technologies, PI, and SCM fields. A Systematic Literature Review identifying more than twelve thousand articles and covering 74 review papers is conducted to identify, select, analyse and synthesise the relevant literature on the integration of disruptive technologies, PI and SCM. The major findings of this paper refer to the identification of the key activities, knowledge areas and strategies in the SC field where the PI and disruptive technologies are gamechanging. These findings are condensed on a conceptual framework that summarises the relationships that exist between relevant disruptive technologies, the PI topics and SC key activities. While a systematic and structured literature review is conducted, it is worth recognising the concerns associated with this paper. The main limitations of this study are attributed to the selection criteria of the documents. First, the use of specific keywords to cover a topic holding a wide variety of knowledge fields. We implemented a forward and backward searchers based on the citations to have a snowballing effect to mitigate this bias risk. Second, this is a tertiary study covering systematic literature reviews. Third, we do not to include conference papers, which might be discussing tendencies, to focus on journal papers and book chapters. The results are limited to the discussion presented in those reviews. Despite these limitations, as the SLR technique is rigorously 14 VOLUME 4, 2016 This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2022.3181154 applied, it is possible to obtain significant knowledge about the research questions [12]. The findings also helped to identify gaps in the literature to be filled by developing new lines of research in the future. From an operational level and considering that organisations are likely to continue developing their analytics capabilities, research might explore the best interactions between data, Operations Research Methods and AI algorithms to optimise and deploy decision-making tools in the SC network. These methods must integrate different forms of data while facing issues of complexity dimensionality, scalability, and interoperability of the problems to be solved. In addition, it is relevant to discuss the adoption of technologies in business cases, and thus consolidate raw data to show the economic benefits on different industrial scenarios. Thus, we expect to see more research on predictive and prescriptive analytics. From a tactical perspective, a research stream might look to define the required abilities, skills, and knowledge of the personnel to keep the new systems running efficiently. Thus, it is important to design strategies to connect professionals with diverse backgrounds by a standardised language around the systems. Other stream might focus on defining protocols and standards to check, measure and control data through inter-organisational SC networks. From a strategical perspective, we foresee some relevant streams. First, analyse the organisational context under which companies achieve success or failure implementing disruptive technologies and moving towards the PI in an early stage. So, case studies are needed to prove key factors to implementation, to set priorities, and to examine cost tradeoffs by analysing various stages of implementation. Second, for companies and governments, it is important to figure out if disruptive technologies are a substitute for the current information systems and decision-making tools or are complimentary of the current ones. Third, it is relevant to show what role the governments play in adopting disruptive technologies, international standards, and protocols. Also, due to real time-sharing data involves all type of partners around the world, it is important to diagnose how the boost obtained by the development of the PI initiative and technological adoption is going to change less developed regions. Lastly, a research stream might be designing and proving new business models into a fully collaborative supply environment and circular economy. .

APPENDIX A
The following This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2022.3181154 VOLUME 4, 2016 This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication.  His research focuses on logistics and quantitative modelling, particularly interested in the resolution of complex, realistic decision-making problems at strategic, tactical, and operational levels in the areas of reverse and sustainable logistics, supply chain management, urban/city logistics, distribution, operations, and scheduling in manufacturing and services. His personal website and email address are https://jrmontoya.wordpress.com/ and jairo.montoya@unisabana.edu.co VOLUME 4, 2016 25 This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2022.3181154