International Transportation | Collection 2022 18 PRODUCTS & SOLUTIONS Logistics Bundling challenges in Hub-and-Spoke networks Focus on rail transport with a case of MegaHub in-Hannover-Lehrte Hub-and-Spoke, Intermodal transport, Rail-Rail transshipment A typical implementation of the Hub-and-Spoke (HS) structure in the rail system is the new rail-rail transshipment terminal, which facilitates the rapid and simultaneous transfer between different traffic flows. Based on literature and focus group interviews, this paper identifies the bundled tasks involved in implementing the HS structure at network and terminal levels and analyzes the potential challenges in terminal operations. Ralf Elbert, Hongjun Wu S hifting freight transport from road to rail and intermodal road-rail transport is an ambitious goal of the European Commission to reduce the greenhouse gas emissions in the transport sector [1]. In this context, Hub-and- Spoke (HS) system has been identified as an efficient way to attract low-volume demands over short distances [2], where the cargos for different origin-destination (OD) pairs could be consolidated to a large and longhaul train in a hub to realize transportation scale economies [3, 4, 5]. A sophisticated bundling between different traffic flows is required to enable the operation of such a complex system and ensure the efficient use of resources [6]. However, it is not an easy task to complete the overall bundling of all relevant traffic flows. In this paper, we will discuss the potential challenges in the bundling tasks of HS structure around a rail-rail and road-rail transshipment terminal, which enables the implementation of the HS structure in a rail system. As a result, we propose answering the following two questions: •• What bundling decisions are involved in- the implementation of the HS structure? •• What are the challenges in operating the rail transshipment terminal? To answer these two questions, this paper will systematically summarize the bundling-related decision problems based on a literature review, and will analyze the challenges in terminal operations of Mega- Hub in Hannover-Lehrte, a typical application of the rail transshipment terminal with hub function, through a focus group interview. Flow bundling around rail transshipment terminals In general, bundling refers to the process of combining traffic flows from different relations into a common train by sharing a part of the journey [7]. It usually happens in 3 basic network structures as illustrated in Figure 1: Line network, feeder/ fork network, and Hub-and-Spoke network [8]. Here in the line network and feeder network, the partly overlapped traffic flows and the traffic flows from the same origin region or to the same destination region, are bundled together in the overlapping paths [9]. In comparison, the center hub connects all directions with long-haul and large-capacity direct trains, each train loads the LU for different destinations and exchanges them with trains to other directions at the hub-[10]. The rail transshipment terminal is a critical connecting point on the rail network, it serves as a central hub connecting the traffic flows in different directions [9, 11]. Distinct from the traditional shunting systems with a hump or flat shunting yard, rail transshipment terminals consist of several parallel tracks to stop trains with loading units (i.e., containers, swap bodies, and semi-trailers) of different destinations, these loading units (LUs) could be transferred to their designated positions of next carrier by sorting facilities, i.e., gantry cranes and shuttle vehicles [6, 8]. According to the involved carriers for LU exchange, the rail transshipment terminals could be divided into rail-rail transshipment terminals and intermodal road-rail transshipment terminals [11]. The former refers to a new rapid transshipment in which the bundled traffic flows (trains) simultaneously appear at the terminal, this operation is also known as pulses or train bundles [6]. Most LUs move directly between trains without intermediate stacking (Figure 2a), so the transfer process could be sped up from one day to a few hours [10]. While most traffic flows via a road-rail transshipment terminal are bundled indirectly as shown in Figure 2b. It means that the exchange process between some traffic flows requires the participation of another subsystem such as the storage system [3] [7], this process always leads to an increasing path and time of movement [12, 13]. The direct exchange only appears when the bundled trucks are waiting at the truck lanes beside the specific container to be operated [8, 11]. Bundling planning on Hub-and- Spoke structure To identify the decision problems related to the planning process for the traffic flow bundling on the HS network and in hubs mentioned previously, a systematic literature review was conducted in the Science Direct and Scopus databases using the search strings (Hub-and-Spoke OR Transshipment) AND (rail OR railway OR intermodal transport). The review results were summarized at the network level and terminal levels. International Transportation | Collection 2022 19 Logistics PRODUCTS & SOLUTIONS Network-level bundling refers to the process of temporal and quantitative bundling of transportation services to ensure cargo flows arrive at their destination on time [7]. This process is always related to the topic “service network design problem (SNDP),” that is, optimally assigning transport services to bundles and optimally assigning cargo flows to services [5]. In many articles, the second step is referred to as a “multi-commodity flow problem” or a “cargo routing problem” [14]. Because there have been very few studies of the SNDP on Huband-Spoke (HS) networks focusing on railrail transshipment, while the majority have focused on liner shipping and combined transport, which could also be applied to the rail network because both have fixed schedules and capacities, the literature, including liner shipping, are also summarized in this section. Time and cost were the main decision factors in the assignment problem, so the minimization of total transit time [15] or total cost (can be transferred from maximization of profit) [2, 16] of LUs was the main optimization objective used in the literature. Some literature used a penalty and discount to describe late and early arrival [17, 18, 19], while some literature decided to cancel the order or increase carriers in the case of overcapacity [5]. The mode of transportation involved heavily influenced the decision strategy. The schedule and capacity of trains were always fixed and served as a hard constraint [15], while truck capacity has been commonly assumed to be unlimited [20], i. e., a truck is always available for an additional service due to its high flexibility, constantly a train is difficult to respond to uncertainties. At the terminal level, the trains via a railrail transshipment terminal have to be assigned to bundle slots of inbound and outbound flows to reduce split moves and revisits, it refers to another topic “transshipment yard scheduling problem” [12]. The number of the vertical and horizontal movements of sorting facilities could also be transferred to a minimization function of cranes’ total processing time [13]. This process was typically constrained by the terminal’s tracks and sorting capacity [12, 13]. Container delays were also considered, with all trains departing according to the fixed schedule and the delayed containers being required to wait at the terminal and be re-assigned to the train in the next circle [12]. Furthermore, reducing the total handling time by optimizing the movement path of LUs is also a trend in the studies for rail transshipment terminals [11]. This process refers to a series of decision problems such as: Determining the positions on outbound trains; assigning the operation area of cranes; determining the handling sequence [13]. Whereas most containers at road-rail transshipments terminal are handled in two spilt moves as shown in Figure 2b [11]. Reducing the handling time in the storage subsystem and increasing the direct move between trucks and trains were usually discussed, which in many studies related to “storage location decision” and “parking position assignment problem” for trucks and trains [11, 13]. Challenges in practical case: MegaHub in Lehrte MegaHub terminal is the first rail-rail transshipment terminal in Germany, which serves as an intermodal terminal for combined road-rail transport at the same time. It was already planned in the 1990s but officially became operational to the public in April 2021 [10]. MegaHub roles as a typical center hub in the Hub-and-Spoke network. As illustrated in Figure 3, five direct train bundles have been in operation via Mega- Hub in Lehrte until the end of 2021 [10, 21]. Each inbound direct train loaded the loading units (LUs) for different destinations and has been operated with a fixed circulation schedule and a fixed number of wagons. The trains meet at MegaHub only in a short bundle slot for LUs’ exchange [21]. A part of the LUs stays on trains and the remainder of the LUs must be transferred to other trains or trucks at MegaHub via semi-automated gantry cranes and Automated Guided Vehicles (AGVs). To identify potential challenges in the actual operations of the MegaHub, we investigated the actual and expected operational conditions of MegaHub using a focus group interview. It was held at the Mega- Hub in Lehrte on May 2021 with MegaHub’s relevant stakeholders, including the terminal operator, the intermodal operator Kombiverkehr and Kombiconsult, and the railway undertaking DB Cargo. The group interview lasted 6 hours and was attended by 3 researchers and 8 invited interview participants with expert knowledge of rail freight transport operations, handling processes, and operations data of the MegaHub. During the group interview, we gained a comprehensive understanding of operational processes and actual operations data in 2021 of the MegaHub from different participants’ perspectives. Some unclear information from the group interviews was clarified via follow-up communication with the participants. After the data collection, we analyzed and summarized the data from the perspectives of flow bundling on the HS network and in the hub and concluded the challenges in the actual operation of the MegaHub as follows. Line network Feeder network Hub-and-spoke network Flow of mixed destination Figure 1: 3 forms of traffic flow bundling on the network All figures by the authors Traffic flow A Traffic flow B TTrraaffffiicc ffllooww AA Traffic flow B Additional subsystem and transport cycle (split move) a. Direct bundling b. Indirect bundling direct move Figure 2: Direct and indirect bundling of traffic flow in hubs International Transportation | Collection 2022 20 PRODUCTS & SOLUTIONS Logistics a. Rail-Road transshipment remains dominant Between 04-12/ 2021, a total of 1,855 trains drove to and from the MegaHub for LU exchanges, involving about 20,000 loading units. Approximately 40% of the LUs were shipped to the bundled terminal without transshipment. Two-thirds of transshipment flows were moved between trains and trucks, with rail-rail transshipment accounting for only one-third. Figure 4 depicts the proportional relationship between these three traffic flows. In terms of data, the MegaHub served more combined road-rail transport in 2021. The role of the central hub function in the rail network was not as significant as expected. This situation is related to the current route planning of trains via Mega- Hub. Intermodal operators stated that bundled trains would be more attractive to customers if they have a shorter transit time or a higher service frequency. However, improving the route of bundled trains remains a challenge. b. Imbalance of inbound and outbound flow Figure 5 depicts the proportionality of inbound and outbound traffic flow in the various bundles. There are apparent differences between inbound and outbound flows in different directions. In connection with Figure 3, an interesting finding about this assignment strategy is that the flow imbalance is more significant for the train bundles with distance imbalance of involved two journeys. The traffic volume between the MegaHub and Verona is more than three times the traffic volume between the MegaHub and Kiel, while the distance between the hub and Verona is about five times the distance to Kiel. Direct trains maintain near-full occupancy over the extra-long distance, lowering the average transportation cost along this section significantly. However, for the relatively short journey, much empty occupancy results in capacity waste and increased average transportation costs. The trade-off between these two parts remains a significant challenge for the flow allocation decision on a rail-based HS network. c. Relative low punctuality The delay problem in German rail freight transport is not new. In 2020, DB Cargo achieved a punctuality rate of 77.6 percent in Germany [22], whereas the punctuality of 5 train bundles via MegaHub is around 60 percent on average in 2021 and varies significantly by direction. The punctuality of direct trains between MegaHub and Rotterdam (and v.v.) is over 85 percent, but only Duisburg MegaHub in Hannover-Lehrte Lübeck Kiel München Verona Rotterdam Lovosice Ludwigshafen Scandinavia Hamburg Figure 3: Train bundles on MegaHub (up to the end of 2021) 41% 20% 39% 0 5000 10000 15000 20000 25000 Loading units Rail-Road transshipment Rail-Rail transshipment No transshipment Figure 4: Composition of loading unit flow operated on MegaHub in 2021 Duisburg DUSS Lübeck Skandinavienkai Hamburg Billwerder München Riem Kiel Verona Interterminal Ludwigshafen KTL Lübeck CTL Rotterdam Europoort Lovosice M e g a H u b 0 5000 Loading units Figure 5: Inbound and outbound flow for different train bundles via the MegaHub (04-12/ 2021) International Transportation | Collection 2022 21 Logistics PRODUCTS & SOLUTIONS around 30 percent between MegaHub and Lübeck Skandinavienkai (and v.v.). Delays can be caused by various factors, including technical failures, constructions, roadblocks, staff shortages, and delays caused by previous delays. Due to critical delays, simultaneous train arrivals are frequently impossible, implying that planned connections are no longer available. In such cases, arriving trains must choose between two scenarios: waiting, which causes further delays, and departing on time, which causes container backlogs at the hubs. d. Semi-automated sorting system MegaHub’s sorting system consists of semiautomated gantry cranes and AGVs between inbound and outbound tracks, with an annual transshipment capacity of 269,000 LUs [23]. Most LUs were directly handled by cranes and AGVs, and some stacking positions also stand in the crane’s operating area for temporary parking. This sorting system is supported by a special operation control system, which aims to minimize energy consumption and wear. It directs AGVs to automatically deliver LUs to the designated wagon, avoiding unnecessary crane movement, and optimizing the operation path and brake process between crane positions to achieve energy efficiency. However, the position selection decision is completed by crane drivers, this operation control system only provides suggestions of open positions for handling. Conclusion Through the literature-based and realworld case studies, we found that the bundling on a rail-based Hub-and-Spoke network revolves around two main points: synchronization and traffic imbalance. The trains on the HS network need to arrive at the hub as synchronized as possible, and the transshipment must be completed in a short time, which puts more pressure on the schedule bundling and the sorting process, as well as makes the trains being more sensitive to time delays. Nevertheless, the amount of rail freight demand varies from region to region, how to balance overloads and empty loads in bundles, and how to stimulate the switch from road to rail remains a challenge. ■ LITERATURE [1] European Commission (2020): Sustainable & smart mobility strategy. [2] Bell, M. G.H.; Liu, X.; Rioult, J.; Angeloudis, P. (2013): A cost-based maritime container assignment model. In: Transportation Research Part B: Methodological 58, pp. 58-70. [3] Alicke, K. (2005): Modeling and optimization of the intermodal terminal Mega Hub. In: Günther, H. O. (Ed.): Container terminals and automated transport systems. Logistics control issues and quantitative decision support. Berlin, Heidelberg: Springer, pp. 307-323. [4] Boysen, N.; Fliedner, M.; Kellner, M. (2010): Determining fixed crane areas in rail-rail transshipment yards. In: Transportation Research Part E: Logistics and Transportation Review 46 (6), pp. 1005-1016. [5] Lium, A.-G.; Crainic, T. G.; Wallace, S. W. (2009): A Study of Demand Stochasticity in Service Network Design. In: Transportation Science 43 (2), pp. 144-157. [6] Boysen, N.; Jaehn, F.; Pesch, E. (2011): Scheduling Freight Trains in Rail-Rail Transshipment Yards. In Transportation Science 45 (2), pp. 199-211. [7] Hu, Q.; Corman, F.; Lodewijks, G. (2015): A Review of Intermodal Rail Freight Bundling Operations. In: Corman, F.; Voß, S.; Negenborn, R. R. (Eds.): Computational Logistics, vol. 9335. Cham: Springer International Publishing (Lecture Notes in Computer Science), pp. 451-463. [8] Kreutzberger, E.; Konings, R. (2016): The challenge of appropriate hub terminal and hub-and-spoke network development for seaports and intermodal rail transport in Europe. In: Research in Transportation Business & Management 19, pp. 83-96. [9] Rupp, J.; Boysen, N.; Briskorn, D. (2022): Optimizing consolidation processes in hubs: The hub-arrival-departure problem. In: European Journal of Operational Research, 298(3), pp. 1051-1066. [10] Kombiverkehr (2022). Transport system MegaHub Hannover Lehrte. Kombiverkehr. https: / / www.kombiverkehr.de/ en/ transport/ megaflexible/ [11] Boysen, Nils; Fliedner, Malte; Jaehn, Florian; Pesch, Erwin (2013): A Survey on Container Processing in Railway Yards. In Transportation Science 47 (3), pp. 312-329. [12] Boysen, N.; Jaehn, Florian; Pesch, Erwin (2012): New bounds and algorithms for the transshipment yard scheduling problem. In: J Sched 15 (4), pp. 499-511. - Fedtke, S.; Boysen, N. (2017): A comparison of different container sorting systems in modern rail-rail transshipment yards. In: Transportation Research Part C: Emerging Technologies 82, pp. 63-87. [13] Cichenski, M.; Jaehn, F.; Pawlak, G.; Pesch, E.; Singh, G.; Blazewicz, J. (2017): An integrated model for the transshipment yard scheduling problem. In: J Sched 20 (1), pp. 57-65. [14] Elbert, R.; Müller, J. P.; Rentschler, J. (2020): Tactical network planning and design in multimodal transportation - A systematic literature review. In: Research in Transportation Business & Management 35, p. 100462. [15] Andersen, J.; Crainic, T. G.; Christiansen, M.: (2009): Service network design with management and coordination of multiple fleets. In: European Journal of Operational Research 193 (2), pp. 377-389. [16] Koza, D. F. (2019): Liner shipping service scheduling and cargo allocation. In: European Journal of Operational Research 275 (3), pp. 897-915. [17] Balakrishnan, A.; Karsten, C. V. (2017): Container shipping service selection and cargo routing with transshipment limits. In: European Journal of Operational Research 263 (2), pp. 652-663. [18] Müller, J. P.; Elbert, R.; Emde, S. (2021): Integrating vehicle routing into intermodal service network design with stochastic transit times. In: EURO Journal on Transportation and Logistics 10, p. 100046. [19] Trivella, A.; Corman, F.; Koza, D. F.; Pisinger, D. (2021): The multi-commodity network flow problem with soft transit time constraints: Application to liner shipping. In: Transportation Research Part E: Logistics and Transportation Review 150, p. 102342. [20] Paias, A.; Ruthmair, M.; Voß, S. (Eds.) (2016): Service and Transfer Selection for Freights in a Synchromodal Network. 7th international conference, ICCL. Lisbon, Portugal, September 7-9. Cham, Switzerland: Springer (Computational logistics). [21] DB Cargo (2021). Was gibt es Neues im MegaHub Lehrte? www. dbcargo.com/ rail-de-de/ logistik-news/ offiziell-im-dienst-megahub-lehrte-6349114 [22] DB (2021). Punctuality. Integrated Interim Report 2021. https: / / zbir. deutschebahn.com/ 2021/ en/ interim-group-management-report/ product-quality-and-digitalization/ punctuality [23] DB AG (2022). We load freight quietly and energy-efficiently. DB. https: / / gruen.deutschebahn.com/ en/ measures/ megahub Ralf Elbert, Prof. Dr. Professor, Chair of Management and Logistics, TU Darmstadt (DE) elbert@log.tu-darmstadt.de Hongjun Wu, M.Sc. Research Associate, Chair of Management and Logistics, TU Darmstadt (DE) wu@log.tu-darmstadt.de Trialog Publishers Verlagsgesellschaft | Schliffkopfstrasse 22 | D-72270 Baiersbronn Tel.: +49 7449 91386.36 | Fax: +49 7449 91386.37 | office@trialog.de | www.trialog-publishers.de Let’s keep in touch editorsdesk@international-transportation.com advertising@international-transportation.com