Physical Internet: First results and next challenges
介绍物理互联网这一物流组织新范式，综述其研究现状、关键问题与挑战，并展示初步仿真结果（如运输里程减少15%、碳排放降低60%），对物流与供应链研究者及从业者有重要参考价值。
The Physical Internet paradigm opens a new way to describe and design how logistics organizations can work, with many managerial, engineering, and economical implications on supply chain performance, including sustainability and resilience. The Physical Internet, as its name suggests, builds on a metaphor from the network of computers networks: the (Digital) Internet. Described in several papers and book chapters (Montreuil 2011; Sarraj et al. 2012; Montreuil et al., 2013; Ballot et al. 2014), its core concept is the universal interconnection of logistics services and networks. To provide an introduction to this special topic forum, we first discuss the origin of the term Physical Internet. Second, we assess the current status of research and provide a literature review. Third, we showcase key PI issues and research challenges and we finish with a brief introduction of the papers that were selected for this special forum. The Physical Internet is inspired by the principles of the Digital Internet, so it is not a mere copy-and-paste of its constituents such as the transmission control and Internet protocols (TCP/IP). This is crucial as there are major differences between data packets on the digital side and parcels and freight on the physical side, and also major differences at the organization levels. The Physical Internet is also by definition different from the Internet of Things (IoT) defined by the connection of physical objects to the Digital Internet. This said, the IoT can be an enabler of the Physical Internet by increasing visibility and control of objects beyond a company’s information systems. The Physical Internet is about interconnecting the world’s logistic networks and is thus defining a new opportunity for supply chain design and operations, enabling seamless open asset sharing and flow consolidation, fulfilling society’s demand for physical objects with an order-of-magnitude better efficiency and sustainability, thanks to improved economies of scale and scope. Physical Internet success stems from interconnecting logistics actors on multiple layers, such as physical, digital, operational, transactional, and legal. Ultimately, the Physical Internet will enable universal interconnectivity with any organization, anytime and anywhere. This is a disruption of the mostly service or customer dedicated logistic networks. At supply chain design and management levels, the Physical Internet opens the way to completely new interconnected operations and business models with a clear goal to improve sustainability in a broad sense. For example, PI implies a redesign of freight transportation, with gradual shift to interconnected transportation. At the basic level, the PI interface will simplify switching between transport carriers (e.g., trailers, railcars) and transport containers. The containers will be moving in a quasi-continuous flow, without driving time limitations or vehicle recharging constraints, and with a continuous tracking of performance and liability to ensure the highest level of service and trust. The impact of the use of PI interfaces for transport containerization on handling and efficiency is well described (Levison, 2016). At a secondary level, PI redesign will improve shipment confidentiality and modular handling containers will improve intercarrier exchanges performed at multiparty sorting and crossdocking hubs, enabling a higher critical mass of flows between hubs, and therefore offer higher transport frequency and higher levels of services (Montreuil et al., 2016). Based on actual data from the consumer goods supply chain, an early simulation-based assessment study of the Physical Internet potential revealed that interconnected transportation enabled decrease of 15% in traveled km, an increase of 33% fill rate, and a decrease of 60% CO2 emissions (Sarraj et. al 2014). A similar transformational shift toward interconnected distribution is achievable by applying Physical Internet concepts to the dynamic smart deployment of goods in an open network of warehouses, distribution centers, and fulfillment centers. Early optimization and simulation-based assessment studies of interconnected distribution revealed significant improvement in efficiency (30% order of magnitude), responsiveness, resilience, and security, through a dynamic network approach securing supplies without duplication of safety stocks and fast fulfillment in line with market expectations (Sohrabi et al. 2016; Yang et al. 2017). The impact of COVID-19 has put a spotlight on such works for all sectors and not limited at the company level like previous analyses (Simchi-Levi et al., 2014). As a new paradigm, the Physical Internet induces changes in logistics organizations and in supply chain applications, but it is also evolving based on trends and supported by new and future research. The Digital Internet was also quite an original paradigm in organizations. Based on a set of protocols, not ISO standards, it was mainly developed by researchers with an associative, thus private governance and gradually adopted by the industry at large, toward its current extensive use across all societal and economic realms. The Digital Internet burst was a disruption compared with the classical interconnection rules already in place between telecom companies in charge of communication in a highly regulated environment. In general companies, and especially the services providers and network infrastructure operators, found in digital Internet concepts, principles, and protocols, notably TCP/IP, the technical solutions needed to settle new businesses with models such as transit contracts and peering bilateral agreements. In short, the Digital Internet brought three main components: a set of protocols independent of technologies, a business framework, and a mostly state-independent governance body. Logistics organizations have different origins. Among these, one is very similar to telecom: the postal services already interconnected under the Universal Postal Union regulations since the end of the nineteen century (https://www.upu.int/en/Home/). This organization still operates but is highly dependent on state-owned operators, sometimes hostage to political stakes, and it has offered few innovations in the last few decades. The other activities remain in the hands of logistics service providers with limited regulations and a continuous flow of innovations in services. To illustrate what PI can provide to the logistic sector, it is useful to consider the same three main interconnection components as previously discussed. From a technical point of view, standardization of tools and processes are not well adopted in the logistics sector. Notable exceptions are the maritime containers on the physical level and incoterms on the transactional level. There is a set of standardized dimensions for cardboard boxes [ISO 3394:2012] yet major players use their own designs. Even for pallets, there exist many standardized sizes, not to mention materials and strengths. The same goes for electronic data exchange (EDI), as messages are standardized but all companies use them in different ways, with minimal intercompany compatibility. The lack of universally adopted tools and processes is a strong barrier against shared solutions and a more efficient logistics process. From a business point of view, a classical approach to develop a logistics business is the expansion of a company by acquiring or integrating competitors in other territories or with specific complementary services. This approach is still at play between logisticians (Carbone and Stone, 2005) and also in the e-commerce sector with companies seeking the integration of logistics companies to maximize their value chain. With the integration, the working methods, the tools, and the codes are defined for the integrating company’s organization which can thus potentially achieve a high degree of consistency, but which remains limited to each such company. Despite the advantages of integration provided by economies of scale and scope, it is limited by investment capacity and antitrust regulations. The second classical approach to develop a logistics business is through the market. Contracting or subcontracting is important in logistics markets, notably for storage, trucking, and last-mile delivery. In most cases, each contract specifies its own set of terms, conditions, tools, and processes. This approach is also very dynamic with the proliferation of marketplaces to ease subcontracting at a larger scale. Between market and integration, a third approach has grown in the last few years, based on collaborative solutions such as alliances, traffic exchange agreements, and pooling (Cruijssen et al. 2007). This approach is the most similar to the Physical Internet paradigm. It seeks to improve the performance beyond the classical boundaries of firms by sharing resources and operations, but with less uncertainties associated with pure market transactions. However, such collaborative organizations, despite some merits, are limited to a few participants and are quite hard to generalize so far. To avoid any misunderstanding, the interconnected approach should not be positioned between the classical organizational approaches to improve logistics performance. It is not a new collaborative organization that would fall between market and integration in a transaction cost framework (Coase 1937). It is a set of protocols, interfaces, and tools, enabling interconnectivity on massive scale and scope that could drastically change business relations in the logistic sector. From a governance point of view, the goal is making the universal interconnection between logistics networks not only technically feasible and economically profitable, but also acceptable by society and industry. One way to make this all acceptable is to demonstrate that the Physical Internet can work, first at a limited scale with experimentations and businesses, so as to build trust and consensus about its design. If collaboration is needed, it is at the design stage of Physical Internet protocols, interfaces, and tools, when researchers and industry innovators can propose solutions and a roadmap, like the EU SENSE project led by ALICE European Technology Platform [https://cordis.europa.eu/project/id/769967]. Concept proofing, pilot testing, experimentations, and improvements are leading the way toward wide scale adoption. At that point, governance of PI solutions will need to take place to define validated Physical Internet solutions and guide their implementation, adoption and evolution. Physical Internet research is enhancing and extending the scientific foundations; assessing the performance improvement potentiality; bridging the capability gaps, notably through new models, protocols, and designs; and validating feasibility and implementation hurdles, particularly through monitoring pilot projects and analyzing case studies (Pan et al. 2017). Research and innovation in packaging, handling, and transport containerization (Landschützer et al. 2015; Montreuil et al. 2016; Sallez et al. 2016) are gradually leading the way toward designed-for-logistics, smart, connected, and ecofriendly Physical Internet containers (e.g., aeler.com, livingpackets.com, poneragroup.com), notably with high-impact industry and trade agreements facilitating their development and deployment (e.g., Leblanc, 2020). Business model innovations in line with Physical Internet concepts are making headway in the market and prospering, as expected from Montreuil et al. (2013b). Examples abound, such as on-demand warehousing (e.g., flexe.com), open-access fulfillment network services (darkstore.com, sell.amazon.com/fulfillment-by-amazon), open access delivery platforms (e.g., roadie.com), as well as freight and logistics marketplaces and apps (coyote.com, freightera.com, colivri, mixmove.io, uber.com/freight). Several large logistic players are currently investigating whether and how to evolve stepwise toward the Physical Internet for themselves. For example, logistics and delivery service providers such as Americold, SF Express, and UPS have engaged in major PI research projects with Georgia Tech’s Physical Internet Center. With multinational corporations, the first steps are usually started by aiming toward a Physical Intranet interconnecting their multiple internal networks and activities, and then gradually consider more open multiparty approaches. As an example, UPS has invested in Ware2Go, a technology company and platform to match merchant needs with flexible fulfillment, recruiting and certifying warehouses in strategic locations, enabling merchants to position products closer to their customers, leveraging the scope and scale of UPS’s network to provide an integrated delivery solution to improve management of the order-to-delivery experience (UPS, 2018). The growing piecemeal PI exploration and adoption by industry, from startups to established corporations, highlights why research and innovation projects with collaboration between industry and academia are so important in the current context. There have been several articles that provided a good systematic literature review of the latest published research in the Physical Internet such as Pan, Ballot, Huang, and Montreuil (2017), Sternberg and Norrman (2017), Matusiewicz et al. (2020) and Treiblmaier, Mirkovski, Lowry, and Zacharia (2020). The following review of recently published PI research provides an update and brief overview of the articles published in 2019 and 2020 that have not been previously reviewed. They also help to position the PI paradigm, identity enablers, and propose implementations with tools or in specific areas. The positioning of the Physical Internet as a new paradigm is still an active scientific debate with several new contributions since last year. Through their literature review, Fergani et al. 2019 propose a general taxonomy for PI, distinguishing between research areas that are not as well covered and providing avenues for further research. Two other papers chose to position PI in comparison with actual approaches. Cornejo et al. (2020) provide an overview of both PI and Lean to show the relationship between both paradigms, and they highlight the potential benefit of value stream mapping for contrasting current and Physical Internet solutions in terms of PI goals. Ambra et al. (2019) exposed the relationships between the concepts of synchromodal transport systems and the Physical Internet, as both were developed to improve socioeconomic conditions and environmental sustainability. Their research identifies potential synergies, future research directions, and critical questions to be considered. Another set of papers focuses on enablers such as the one proposed by Meyer et al. (2019). It develops a Blockchain-based 4-layered framework to overcome some of the barriers within PI associated with the exchange of value and physical assets in decentralized logistics networks. Betti et al. (2019a, 2019b) investigate the exploitation of Blockchain distributed ledgers and smart contracts in interconnected logistics and validate the potential by coupling an agent-oriented discrete-events simulation with a Blockchain platform. In the same vein, Tran-Dang et al. (2020) investigate the application of Internet of Things technologies, building blocks, and a service-oriented architecture to accelerate the implementation of PI. Propose an open network-model approach for providing infrastructural data sovereignty that will enable the sharing of sensitive operational data as required for realizing PI. From another perspective, Lafkihi et al. (2019) use gamification methodology to compare a centralized approach, based on a central authority that optimizes transport plans for all carriers, versus a decentralized approach where carriers optimize their own transport plans, as found in simple PI implementations. Results indicate centralization outperforms in terms of global efficiency and effectiveness; while decentralization is better for individual incentives. The last proposed set of papers focuses on solutions for existing problems or new problems raised by new types of operations. Osmólski et al. (2019) present dedicated PI solutions to logistic processes such as modular transport units and real-time planning and information exchange, as well as properly communication et al. (2019) the use of interconnected and systems for an existing freight in a PI leading to a dynamic real-time for et al. (2019) a optimization model that can be for dynamic and within the industry. et al. (2019) focuses on operations in a They the as a model with and validated through an et al. (2019) a simulation that multiple in a flexible dynamic can a shift toward transport that are useful in PI. et al. (2019) a key of the Physical Internet is the need for interconnected that a and they that a model better for high vehicle and a model better for vehicle From a perspective, et al. (2020) the in leveraging PI as a strategic development and by a strategic for to ensure its place as the most logistics by There were papers to the original special topic for was by a of three at each with some papers through leading to papers as The of this special topic as on the Physical Internet was it provided an opportunity for researchers to the latest technologies, applications, and to the Physical Internet. Second, it to critical issues and challenges for future research and development in the broad of Physical Internet and of interconnection and of logistics networks and supply Third, it to further logistics research the new Physical Internet paradigm. the three papers selected for this special goals. The first from and the Digital Internet to the Physical A framework with a network and the between Digital Internet and the Physical Internet. This a framework for PI based on the Digital Internet with the of both the and It the of and the of PI in comparison with In the propose a network model to the implementation of the PI. the develop an to the model and demonstrate how it can be to the PI in a case The second from and is in a Physical Internet and et al. It the and success associated with in a PI network both and They use a research approach, and three logistics service providers in a to demonstrate that central and of resources is a and in PI, especially with continuous PI The third from Sternberg and the Physical Internet Logistics service and design and a key of the PI. The of PI containers is the of this as it at the design and that will the flows in a network context. They a model that flow to investigate compatibility. The that in terms of and flows whether PI or compared with the existing logistics The also show the of and on what the of technology which are all important for future research on and design. The from and A of within a in and the concept of an open network that can access to place for that will in existing to a simulation the were to demonstrate the of closer to leading to in private and existing service providers with only in to existing freight The show more more efficient of leading to a more to the company’s the Physical Internet a paradigm for analyzing logistics operations that particularly to the needs of the environmental and and resilience. It therefore that companies not to the associated Even the first of implementation we remain from universal including when technical solutions are with associated economic for modular handling containers Several which are not can be put which the many avenues of research that are to closer to a implementation of the Physical Internet. There is a need for more operational of PI, both on a larger scale and more open to in the research some companies use concepts to the Physical Internet, in Physical Intranet the limited communication on this the of still The open-access sharing of resources and flow consolidation, in the broad between logistic companies should therefore to be the of research to better the associated with their performance and the conditions for their The of operations is such that to and will also be to the for each of the operators, as by Lafkihi et al. (2020). The distribution of a logistic service several including not at the and beyond is a strong point but the of trust and for monitoring performance. the needs for shared and a or a to a shipment to a third with a to that of operations to a The of the maritime 2016) the of a logistics but all of it on a global scale. the and and it is to in between the players in order to achieve a shared was (e.g., has been and to a new to containers without any company. the design of the remains an and in a limited way the issues at to a large and sector in a to logistics efficiency with the of its This is the by the European which a with industry to this In this it also be useful to from other sectors such as and its Technology for For years, this sector has been and more which is not a physical but a business in the of logistics a of in efficiency set by could make it to technical As the paradigm of the Physical Internet many and but research questions and we are that this special will significant to be on the Physical Internet through the research it will are of all the their the the to the articles found in this special and especially the of Business Logistics and their and providing with the opportunity to develop this