Transportation Research Part E: Logistics and Transportation Review
Hinterland freight transportation replanning model under the framework of synchromodality
Introduction
Hinterland freight transportation is carried out in accordance with an elaborated tactical plan. A service Network Design (SND) problem lies at the core of the tactical planning process, addressing issues like the selection and scheduling of services, specification of terminal operations and routing of shipment flows (Crainic, 2000). These plans are designed on the basis of predicted freight volume. Stochastic and robust planning are studied extensively to prevent or limit the influence of prediction bias.
At the operational level, the tactical schedules are implemented on a local scale with a shorter time window. During this process, the transportation network is vulnerable to uncertainties and variations. Those not only rise from the external environment (unexpected additional or cancelled freight demands, weather caused hazards, delays of the shipment release), but can also internal from system fluctuations (congestion, breakdown and human-caused malpractices). The stochastic or robust planning plan is unable to handle all unexpected scenarios in everyday practice. It is still on open filed to regulate uncertainties at the operational level.
Synchromodal transportation is an upcoming new solution succeeding intermodal transportation towards a more flexible and cooperated freight transportation (Tavasszy et al., 2010) at operational level to deal with uncertainties. Under the framework of synchromodality, shippers accept a mode-free booking and only determine the price and quality requirements (Behdani et al., 2016). The involved service providers collaborate under the centralized supervision of one Logistic Service Provider (LSP) (see also SteadieSeifi et al., 2014, Li et al., 2017, Pérez and Mes, 2017). This implies that it is possible to perform real-time switching among different modes (Behdani et al., 2016).
The centralized LSP holds the promise of generating specific replanning solutions to the specific uncertainties encountered in each individual case. The replanning questions come in: (1) Which part of the original SND plan can be re-planned? (2) To what degree can the re-planning solutions deviate from the SND plan? As mentioned above, the SND mainly consists of three components: the shipment route choices, the terminal operations, and the service schedules. With regard to the first component, it is important to deliver shipments to their destinations on time. Hence, complete flexibility can be given to the shipment route replanning. This is under the agreement of mode-free booking and real-time switching. The thorough re-routing of shipments unavoidably involves the replanning of terminals operations and services schedules. Those can be integrated in the replanning and benefit from the cooperation of all synchromodality operators.
This replanning model further carries out two consecutive synchronization tasks: to derive the corresponding transshipment flows at the terminals, and to synchronize the service rescheduling with the shipment rerouting. A dilemma emerges for to the LSP while rescheduling the LCS. The higher degree of flexibility given, the more flows can be diverted from truck services to LCS. However, that will also generate higher costs as a result of deviating from the SND plan (paying for cancellations or postponing services) and retrospective re-positions. The trade-off between flexibility and deviations is crucial for the rescheduling of LCS. This work explores how LSP can synchronize these intertwined tasks for hinterland freight transportation replanning.
This study includes the following contributions to existing research. We construct a model that regulates the shipment flow re-routing and services re-scheduling. It can be used for a hinterland freight transportation network under the framework of synchromodality. The shipment rerouting is given a higher degrees of freedom, involving the split and bundling of shipments. Consecutively, the transshipment flows at the intermediate terminals are extracted from the shipment flow rerouting. Transshipment is an important link in the overall process, albeit one that is sometimes neglected and over-simplified in some studies. Last but not least, moderate flexibility is given to the operating times of services, to be synchronized with the shipment flows. Alignment is introduced to LCS, i.e., the barge and rail services are given a buffer time that transcends the scheduled time. Thus shipment flows en route can be shifted from truck services towards LCS.
The model can be used for daily planning at the operational level. It can be embedded as the ‘real-time layer’ of the SYNCHRO-NET service platform (see Fig. 2 of Giusti et al., 2019). In this paper, the model is tested in the Rotterdam hinterland transportation network, in cases involving late release of shipments, latency of LCS, volume fluctuation of shipments, and a variety of mixed perturbations. The model can provide a replanning resolution within seconds. In addition, we compare this model to other rigid replanning methods. The results show that the proposed synchromodality replanning model can save more overall operating costs, improves the modal split of barge and rail services, and improves LCS utilization. Finally, the model is tested with theoretical instances, with different network sizes, shipment amount and network typologies, providing that the model is able to solve small-size problems efficiently.
The remainder of this paper is organized as follows. The next section provides a review of relevant studies those incorporate uncertainties in transportation planning/replanning. The problem is formally described in Section 3, and presented mathematically in Section 4. The mathematical model describes the synchronization scheme of shipment re-routing, transshipment flows at the intermediate terminals and service rescheduling. The replanning algorithm is calibrated and described in Section 5. Section 6 tests the proposed model in the Rotterdam hinterland network with a variety of perturbations, with the solutions being provided tailored to the specific situations. The model is further analyzed in Section 7 in three aspects: (1) critical path and re-scheduling flexibility; (2) comparison with other two rigid replanning methods; (3) theoretical instances to check the solving difficulties. Finally, in Section 8 we present our conclusions and discuss possible avenues for future research directions.
Section snippets
Literature review
In this section, we take a look at the studies which deal with unexpected fluctuations in freight transportation planning. The studies discussed here can be divided on the basis of the uncertainties (also called perturbations) involved.
Uncertainties in freight transportation can rise in exogenous and endogenous form. Exogenous uncertainties mainly come from the shipments, in the form of volume fluctuations, release date deviations, and the corresponding mutual impact. Endogenous uncertainties
Problem description and assumption
Consider a synchromodal network engaged in transporting shipments (or freights) from many origins to many destinations, using various service modes. Our model implements the re-planning of involved operators to cope with uncertainties at the operational level.
Let N denote the set of nodes of the network, standing for terminals which can be ports, railway stations or truck hubs. Mark each service by a sequence number . The arc (also named as link or leg) indicates a service v from
Mathematical formulation
The aim of this section is to construct an integrated and detailed model of the shipment flow assignment replanning and real-time deployment of the service fleet. We model the online replanning in three parts. (1) the shipment flow re-planning for a capacitated network denoting the transportation of shipments, at the arc level; (2) The service rescheduling at the arc level, mainly involving the service rescheduling and service usage. (3) The shipment flow re-planning and service rescheduling
Re-planning algorithm
The general steps of the replanning procedure are included in Algorithm 1: Algorithm 1 Re-planning Algorithm 1: Input: Set of unexpected events, current status of the network. 2: Output: Re-planning solution, the integrated shipment flow rerouting, transshipment and service rescheduling. 3: procedureProcedure re-plan when (an) unexpected event(s) occur(s) 4: Step 1: Initialize the network. 5: Step 2: Apply the pre-processing: 6: Step 3: Solve the STP model.
Computational experiments
In this section, we demonstrate the applicability of the model in a real-word transportation case. The optimization model is coded and solved in CPLEX 12.8.0 on a computer with an Intel Core i5-4690 3.2 GHz processor and 8 GB of RAM. All the instances are solved within the time limit within seconds.
The computational experiments are performed in two parts. Firstly, we present a description of the Rotterdam hinterland network and its daily transportation tasks. A base case is provided as an
Computational analysis
Based on the computational experiments, we also analyze it looking at three other aspects. Firstly, we check the critical path of the example network to see if it is possible to improve the service rescheduling flexibility any further. Secondly, the proposed model is compared to other rigid re-planning methods to access the value of shipment split and flexible service scheduling. Finally, the model is tested based on more theoretical instances to explore the difficulty of the model resolution.
Conclusions and future research
In this paper, we propose a synchromodality transportation replanning (STP) model for a hinterland freight transportation network. To this end we present a mixed-integer linear programming model, determining the minimum cost solution by combining decisions integrating shipment rerouting and service rescheduling. We demonstrated the performance of the approach in the case of the Rotterdam hinterland transportation network. The STP model provides specific resolutions according to the specific
Acknowledgement
The first author gratefully acknowledges support by the China Scholarship Council under Grant 201407090053.
References (30)
- et al.
Stochastic service network design with rerouting
Transp. Res. Part B: Methodol.
(2014) - et al.
The multi-objective railway timetable rescheduling problem
Transp. Res. Part C: Emerg. Technol.
(2017) - et al.
Modelling network travel time reliability under stochastic demand
Transp. Res. Part B: Methodol.
(2005) - et al.
Optimal multi-class rescheduling of railway traffic
J. Rail Transport Plann. Manage.
(2011) - et al.
Integrating train scheduling and delay management in real-time railway traffic control
Transp. Res. Part E: Logist. Transp. Rev.
(2017) Service network design in freight transportation
Eur. J. Oper. Res.
(2000)- et al.
Scenario grouping in a progressive hedging-based meta-heuristic for stochastic network design
Comput. Oper. Res.
(2014) - et al.
Dynamic approach to solve the daily drayage problem with transit time uncertainty
Comput. Ind.
(2013) - et al.
A tabu search algorithm to solve the integrated planning of container on an inter-terminal network connected with a hinterland rail network
Transp. Res. Part C: Emer
(2018) - et al.
Intermodal freight transport planning– a receding horizon control approach
Transp. Res. Part C: Emerg. Technol.
(2015)