A simulation to examine a synchromodal hinterland transportation network

(1)

Master’s Thesis

A simulation to examine a

synchromodal hinterland

transportation network

Bram de Wilde s1916173 Supervisor: Dr. E. Ursavas Co-assessor: Dr. K. Scholten

Supervisor company: ir. B. van Riessen

MSc Supply Chain Management

MSc Technology & Operations Management Faculty of Economics and Business

(2)

2

Abstract

(3)

3

Preface

Firstly, I would like to thank ECT and EGS for giving me the opportunity to write my master’s thesis on such an interesting, important and innovative topic. In particular, I would like to thank my supervisor of ECT, Bart van Riessen. His extensive knowledge on the topic has guided me during the project. Feedback and discussions always led to new insights. I would also like to thank Joost Koning and Paul Zoeter for providing me with essential information on the network.

Secondly, I would like to thank my supervisor Dr. Ursavas. She brought me into contact with ECT and her enthusiasm about the research was of great support during the project.

(6)

6

1. Introduction

1.1 Introduction

Europe is experiencing a rapid growth in freight transportation due to recovery of the financial and economic crisis and an expanded European Union (Eurostat, 2012). Managing these increased flows will be a serious challenge. In order to stay competitive as a continent and to prevent further environmental damage, existing infrastructure should be used more efficiently.

Ports are important links in these growing freight flows of today's complex supply chains, since a significant part of freight transportation occurs from or to a port. Earlier paradigms viewed ports only as a distinct entity with a function in the value chain (Robinson, 2002). However, due to increased complexity and dynamics of port operations, they should not be viewed in isolation anymore (Robinson, 2002). Integrating port operations in the supply chain is therefore of utmost importance, since chain systems compete with chain systems in the globalized economy.

After liquid and dry bulk, container transportation accounts for the largest weight of freight transportation (746.5 million tons) in Europe (Eurostat, 2014). The importance of container transportation and the proportion that it holds in the transportation market have increased significantly over the last two decades (Zhang et al., 2009). Container terminals in ports handle millions of TEUs (Twenty-foot Equivalent Units) every year and serve as a global hub and gateway for many deep-sea liner services. Besides accessibility and efficiency of the loading and unloading of containers within the terminal, increasing emphasis is being paid to the transportation of containers to the hinterland.

(7)

7

Horst and De Langen (2004) state that hinterland access is an important issue in ports, since it is a key success factor of European ports.

Intermodal transportation has been strongly encouraged by the European Union to enhance overall efficiency and to lower the impact on the environment by choosing more sustainable transport modes (European Commission, 1997). Transporting these flows for the largest part by road will lead to significant road congestion in Europe. Furthermore, road transportation has a higher carbon dioxide emission per TEU than rail and water transport. Hence, more TEUs should be transported by rail and water.

A new logistic concept, related to intermodal transportation, is synchromodal transportation. With synchromodal transport, a decision is made for every transport between different modalities (road, rail, water), based on the most efficient and sustainable solution. Synchromodal transportation could also be described as real-time intermodal transportation planning. This offers flexibility for the carrier when e.g. a waterway is blocked for a certain period. This could subsequently decrease transit times, reduce costs and improve reliability for the customer. Furthermore, synchromodal transportation could be the solution to the increasing environmental concerns in Europe by choosing modes of transport that are more sustainable than road transport. Synchromodality is strongly advocated in the Netherlands and is one of the most important issues formulated by the Topteam Logistiek to reach the top position of the World Logistic Performance Index in 2020 (Topteam Logistiek, 2011).

(8)

8

important to be competitive for a network operator, especially in a business-to-business environment where customers demand reliable transportation (Keskinocak & Tayur, 2004). Increasing environmental concerns in Europe require more sustainable transportation in order to prevent further environmental damage. A conceptual study, based on a real-life case, whether synchromodality can be a good solution under expected capacity constraints is not yet performed in literature. This study thus aims at enriching the literature on synchromodal transportation and examining the possible advantages in practice, which makes the study relevant for both theory and practice by means of simulation. The outcomes of the in-depth case-based study contribute to the further development of the concept. Hence, it can lead to new insights, e.g. which aspects of synchromodality are most promising and which seem to be less important. Besides the outcomes of the experiments, the simulation model itself can be used as a tool for examining other cases.

1.2 Research questions

To enrich synchromodal literature and examine the different aspects and possible advantages of synchromodal transportation, this study will answer the following research question.

Can the concept of synchromodal transportation solve disruptions in a hinterland transportation network?

(9)

9

2. Theory and background

2.1 Introduction

In this section, relevant literature on intermodal transportation is reviewed first to give a brief overview on the concept, since it forms the basis of synchromodal transportation. Secondly, publications on the concept of synchromodal transportation are covered. Important aspects of synchromodal transportation that will form the basis for the experimental design will be introduced.

2.2 Intermodal transportation

Intermodal freight transport is the term used to describe the movement of goods in one and the same loading unit or vehicle which uses successive, various modes of transport (road, rail, water) without any handling of the goods themselves during transfers between modes (Macharis & Bontekoning, 2004). A decision thus has to be made within the container terminal which mode of transport will be used to transport the container to the hinterland, hereby not handling the content itself. Intermodal transportation is strongly advocated by the European Union in the 1990s. The European Commission (1997) stressed the importance of a more efficient use of the transport system by stating that it is an essential prerequisite for the competitiveness of Europe, especially when freight transport would increase when Central and East European countries will join the European Union, which is now the case. Integration of different transport modes in a door-to-door transport chain, thereby exploiting the favorable characteristics of each mode, can lead to shorter lead times and lower costs (e.g. by avoiding road congestion). Besides this economic incentive, environmental concerns have driven the support for intermodal transportation.

Literature emphasizes the importance of lowering emissions and argues that the choice of mode for hinterland transportation can significantly influence the amount of emitted CO2.

Lättilä et al. (2013) argue that dry port (an inland port connected with a sea port by rail and barge connections) usage can both decrease emissions and costs by using rail between the dry port and the sea port instead of road. Craig et al. (2013) calculate a 46% lower amount of CO2

(10)

10

road still accounted for 74.9% of the total inland freight transportation in 2011 (expressed in tonne-kilometres). Rail and inland waterways accounted for respectively only 18.8% and 6.3%. This suggests that rail and barge connections have important drawbacks compared to road transport. Boysen et al. (2013) review freight traffic transported by rail, which is in steady decline in spite of substantial support programs of the European Union. Despite the advantages of avoiding congested roads and less environmental derogation, they state that the delivery speed of a freight train is somewhere between 10 km/h and 20 km/h, mainly caused by long waiting times at rail yards. Besides this low speed, reliability is also low. Only 53% of freight trains arrive with less than half an hour delay (Boysen et al., 2013). Konings (2007) states that barge connections have a strong position in the port of Rotterdam. However, transit times of barge services are relatively long due to inefficiencies in handling barges. Barges have to call at many terminals in the port, which causes unnecessary waiting that could be used more productively, e.g. sailing (Konings, 2007).

Intermodal transport is thus well covered in literature. However, it mainly focuses on similar loading units and not on the efficient use of transportation networks. SteadiSeifi et al. (2014) describe synchromodal transportation as the next step of intermodal transportation. Intermodal transportation thus forms the foundation for synchromodal transportation.

2.3 Synchromodal transportation

2.3.1 Synchromodal transportation defined

(11)

11

the loading capacity of road trucks, rail shuttles or barges. This is in line with the definition of Van Riessen (2013): synchromodality is an intermodal transportation network with online planning. SteadiSeifi et al. (2014) also emphasize the flexibility aspect of synchromodality. Synchromodal transportation can thus be seen as real-time intermodal transportation or as flexible intermodal transportation. Decisions are not fixed, but are subject to changes such that the most efficient and sustainable solution can be chosen for every trip, taking into account current conditions. This flexibility is the key difference between intermodal transportation and synchromodal transportation. The definition of Van Riessen (2013) will be used in this study as the definition of synchromodal transportation.

2.3.2 Potential of synchromodal transport

(12)

12

Synchromodal transportation can thus make transportation networks more effective through coordination between different parties, which could lead to lower costs, less greenhouse gas emissions and improved efficiency.

Finally, synchromodality can be the solution for achieving the modal split target set by the Port of Rotterdam. The modal split between the truck, barge and rail was respectively 44.3 %, 42.9 % and 12.8 % in 2012 (Port of Rotterdam, 2013). The modal split target set by the authorities for 2035 is 35 %, 45 % and 20 % for truck, barge and rail respectively. A more flexible use of the different modalities can make the target more feasible.

2.3.3 State of the art synchromodal transportation

Publications on synchromodal transportation are relatively scarce. Synchromodal transportation is almost not covered in scientific literature. However, it is likely that this will change, since the concept looks promising and it is strongly advocated by important stakeholders.

(13)

13

based on mutual trust and partnership is essential in synchromodal transportation. Furthermore, the significance of quality features of unimodal and synchromodal transportation are compared based on questionnaires filled in by companies that actively participate in organizing freight flows. Results show that synchromodal transportation necessitates a high involvement of all parties compared to unimodal transport.

Publications on synchromodal transport thus emphasize the need for cooperation, co-ordination and information sharing. Furthermore, there exist mathematical papers on the planning of synchromodal transport. A conceptual study, based on a real-life case, whether synchromodality can be beneficial under expected capacity constraints is not performed. This study can therefore lead to new insights that can be useful for the further development of the concept of synchromodal transportation.

2.3.4 Dimensions of synchromodal transportation

Van Riessen et al. (2014) state that synchromodal transportation can be split up in three dimensions. These dimensions are three different degrees of flexibility: flexibility in time (due date), flexibility in modality and flexibility in the route chosen. Flexibility in time exists when the due date of e.g. a container can be stretched. It can also be described as due-date tightness (Baker, 1984). Obviously, flexibility in time creates more flexibility in planning, such that capacity can be used more effectively and efficiently. It is expected that this will subsequently lead to a better performance, since the flow allowance (time between release date and due date) is increased (Baker, 1984). Although due dates are set exogenous by customers, there are possibilities in stretching due dates.

When a container may be transported on all available modalities, there is flexibility in modality. Flexibility in modality gives the opportunity to a network operator to optimize the utilization of available barge and train services (Van Riessen et al., 2014). Finally, flexibility of route implies that a container may be routed indirectly, e.g. via another terminal. Routing flexibility is a well-known topic in operations literature that can improve performance (e.g. Das & Nagendra, 1997; Koste & Malhotra, 1999; Zhang et al., 2003). It is especially treated in flexible manufacturing systems. There is a clear link with transportation, since it allows firms to find alternate processing centers in case of machine breakdowns or system overloads (Zhang et al., 2003). Translated to transportation networks, flexibility of route could hence help a network operator in avoiding network disruptions.

(14)

14

different tariffs: a lower tariff for more flexibility and a higher tariff when no flexibility is allowed. This concept should be adopted from the airline industry, where revenue management belongs to its core principles nowadays.

When flexibility in all three dimensions is present, one can speak of theoretically perfect synchromodality. These three dimensions of synchromodal transportation form the basis for the experimental design (see section 4.4 and Figure 2.1) as the main variables of the study, such that can be examined which dimension is most important and which dimension is less important.

Figure 2.1: Dimensions of synchromodality 2.3.5 Key performance indicators of freight transportation

Key Performance Indicators (KPIs) are metrics, both financial and non-financial, which are used for defining and measuring the progress towards the goals set by an organization. In freight transportation, several KPIs exist. Standard performance measures are on-time performance, transit time, the costs of transportation, reliability, flexibility and sustainability. Vetevood (2008) examined the requirements for an evaluation tool for intermodal transportation. Main KPIs identified in the study of Vetevood (2008) are general transport cost, transportation time, flexibility, reliability and accessibility. This is in line with Keskinocak and Tayur (2004), who state that reliability is especially important in a business-to-business environment, which is the case here.

(15)

15

3. Case description

3.1 Background information ECT/EGS

European Gateway Services (EGS) is a subsidiary of Europe Container Terminals (ECT). Europe Container Terminals is the largest container terminal in Europe and achieved a throughput of 7.7 million TEU in 2012 (ECT, 2013). It is founded in 1966 and has about 2300 employees. ECT has three deep-see terminals, the Delta terminal, the City terminal and Euromax terminal, all in Rotterdam. As an additional service, ECT offers its customers access to the hinterland. This service is offered under the name European Gateway Services: an expanding network of inland terminals, e.g. Trimodal Container Terminal Venlo and DeCeTe Duisburger Container terminal. The network of EGS can be found in Figure 3.1. Recently, EGS has expanded its network: Munich, Enns (Austria) and Vienna (Austria) are added. The inland terminals function as extended gate, such that companies can take advantage of additional customs, like paperless transport. With paperless transport, freight travels under the customs license of ECT and only at the extended gate the customs formalities have to be arranged, which can save time and money. EGS offers customers frequent barge and rail connections to the hinterland. On trains, the company has dedicated capacity. Synchromodal transportation is the starting point for EGS in delivering these services. A detailed network description with the different connection with each inland terminal can be found on the website of EGS (EGS, 2014).

(16)

16

3.2 Problem description

As described above, EGS controls an extensive network within the hinterland. An important inland terminal, also extended gate, is the DeCeTe container terminal in Duisburg. EGS has frequent barge and rail connections with the terminal in Duisburg. The most important rail route that connects Rotterdam with Duisburg is the Betuweroute. The Betuweroute starts at the Maasvlakte in Rotterdam and directly connects the Port of Rotterdam with Germany. The Dutch part of the Betuweroute is only available for freight trains. In Germany however, also passenger trains use this railroad. Capacity for freight trains is therefore limited and additional capacity is needed to meet increased freight traffic from the port of Rotterdam in the near future.

To enlarge capacity, a third track will be developed between Emmerich and Oberhausen between 2015 and 2022. Severe disruptions are expected on this part of the Betuweroute during the development of the new track. Capacity in the long-run will thus be enlarged, however while constructing the additional track, capacity will be limited. In Figure 3.2, the problem situation is depicted schematically. The red line indicates where disruptions will occur. The expected disruptions can be categorized in four different week types (Table 3.1). Currently, between 500 and 550 trains per week travel on the Betuweroute (Keyrail, 2015). As can be seen in Table 3.1, in some weeks (week 2), capacity is reduced to 280 trains per week, which is almost a decrease of 50 % compared with the current number of trains that use the Betuweroute.

(17)

17 Week Start disruption End disruption Capacity (trains/day) How many weeks? 1 Monday Friday 40 17 2 Monday Sunday 40 98 3 Friday Sunday 0 36 4 Friday Sunday 25 144

Table 3.1: Overview disruptions Betuweroute

To ensure sufficient capacity to Duisburg and prevent severe congestions, EGS should consider several logistic alternatives. Different options exist, such as alternative rail routes and synchromodality. As can be seen in Figure 3.2, ECT has an inland terminal near Duisburg, the Trimodal Container Terminal in Venlo. Possible solutions to avoid the Betuweroute are therefore to reroute containers via Venlo by train (Brabantroute) to Duisburg or employing more direct barge connections between Rotterdam-Duisburg. Possible solutions will be described more specifically in the methodology.

(18)

18

4. Methodology

4.1 Simulation

The research provides a case-based simulation study within European Gateway Services. A simulation model is used to test whether synchromodal transportation could be a good alternative in solving capacity constraints compared to other logistic alternatives. Simulation is chosen, since it can cope with variability, interconnectedness and complexity, which will all be present in this case (Robinson, 2004:4). Furthermore, in simulation, experimental conditions can be controlled easily, such that a direct comparison between alternatives is possible. Finally, simulation provides transparency for stakeholders, compared with e.g. mathematical modeling (Robinson, 2004:9). It is intuitive and it will be more convincing in communicating the outcomes of the study. Simulation has however also some limitations. Model development is time consuming and there is a need for much data.

The model is built by using Tecnomatix Plant Simulation 10.1. Plant Simulation uses a discrete-event simulation approach and is suitable for all kinds of logistic systems (e.g. production logistics, transport logistics and storage logistics). Plant Simulation uses SimTalk as a coding language and is suitable for experimenting with different parameters and alternatives. Common random numbers can be used in Plant Simulation during experimentation. With common random numbers, generated streams of numbers between experiments are similar, but within an experiment (replications) the numbers differ. This makes debugging and comparison between experiments easier. Finally, the software is able to present output in a logic way and exporting to e.g. Microsoft Excel is possible.

4.2 Model explanation

The model will examine the triangle of terminals of ECT: Rotterdam, Duisburg and Venlo, which can also be seen in Figure 3.2. Only import containers are considered, since the numbers of import containers are larger than the number of export containers. Containers hence arrive in Rotterdam. The time that containers are available for transport is taken as the arrival time, and not the arrival times of the container in the port. Only the time available for transport is relevant for this study, since the actual operations on the terminal are out of scope.

(19)

19

Austria are routed on separate trains via Venlo over the Brabantroute and are therefore taken into account. Different user-defined attributes are assigned to containers, like Final Destination, ContainerID, ArrivalTime and DueDate. The number of containers that arrives, the distribution between the destinations, arrival times and due dates are based on detailed data that was available.

Different modalities (transporters) are created according to the time schedule of EGS. Barges and trains are created two hours before departure, such that they can be loaded and subsequently can leave at the right time. Trucks are created when there are more than two containers waiting in the buffer in front of the truck loading. To calculate the costs, the costs of the trips to the destinations are built in the model. Finally, the capacities of the different modalities are based on detailed data. Data of the modalities and the containers are registered at the end of the simulation.

4.3 Assumptions

Since not all complexity of the problem can be captured in the model, assumptions are made to simplify the situation and the focus on the most important aspects of the problem situation. The most important assumptions are listed below. Assumptions are based on data of the company and discussions with employees.

 The three terminals in Rotterdam are merged and taken as one terminal.

 Trucks arrive when there are more than two containers in a truck buffer.

 The speed of a barge is normally distributed with an average of 3.9 m/s and a standard deviation of 0.5 m/s.

 The speed of a truck is normally distributed with an average of 19 m/s and a standard deviation of 1.4 m/s.

 The speed of a train is normally distributed with an average of 13 m/s and a standard deviation of 0.5 m/s.

 Containers with final destination Duisburg and a 'time till due date' larger than 2.5 days can be transported on all three modalities.

 Container with final destination Duisburg and a 'time till due date' larger than 1 day and smaller than 2.5 days can be transported on only the train and a truck.

(20)

20

 Containers with final destination Venlo and a 'time till due date' larger than 1.5 days can be transported on all three modalities.

 Container with final destination Venlo and a 'time till due date' larger than 0.75 day and smaller than 1.5 days can be transported on only the train and a truck.

 Containers with final destination Venlo a 'time till due date' smaller than 0.75 day are only allowed to be transported on a truck.

 Roads and waterways do not have failures (i.e. traffic jams, high water).

 Passenger trains on the Brabantroute are not modeled. Only capacity assigned to freight trains is taken into account.

4.4 Input parameters

An overview of the input parameters of the model can be found below in Table 4.1.

Container (Inter)arrival times Distribution due dates Distribution final destination

Costs Trips Per modality Per destination

Trips Barge, train Speed Capacity

Time schedules EGS Distance to destination

Truck Capacity

Speed

Distance to destination

Network Time schedule Betuweroute

Time schedule Brabantroute freight trains

Disruptions Betuweroute (third track development) Disruptions Betuweroute (copper theft, accidents etc.)

Disruptions Brabantroute (copper theft, accidents etc.)

(21)

21

4.5 Model verification and validation

Verification is the process of ensuring that the model design has been transformed into a computer model with sufficient accuracy (Robinson, 2004: 209). In this study, the main objective is to examine the three different dimensions of synchromodality (see 4.7 Experimental design). Verification should thus ensure that these three dimensions are present in the model such that they can be examined alone and in combination. This is ensured by micro checks throughout the development of the model by the modeler. Furthermore, discussions with stakeholders and experts have clarified the dimensions, such that it could be translated to the model. In this phase, some containers did not arrive at their destination. With the help of the common random numbers in Plant Simulation, these containers could then be traced to detect the error in the model.

Validation, on the other hand, is the process of ensuring that the model is sufficiently accurate for the purpose at hand (Robinson, 2004:210). The model should examine the different scenarios on especially on-time performance and costs. These performance measures are identified as most important by means of discussions with different stakeholders. A container is on time when it arrives at the terminal before the moment a customer needs the container. On-time performance could be interpreted as reliability in the study of Vetevood (2008). The costs will be calculated as the sum of the costs of all trips performed. The model will only be accurate if other parts of the model are accurate as well. Therefore, e.g. transit times of transporters, the routing of containers and the due dates of containers in the model are compared frequently with real data. Furthermore, most important codes are checked before experimenting and visual checks are performed. Subsequently, initial output reports are inspected. Finally, initial results, like utilizations of different modalities, are discussed with 'experts' within the company. These discussions have led to small alterations of the model, such that final results are accurate.

4.6 Experimental design

(22)

22

and hence using the same modality should solve the expected problems. Since the Brabantroute is a very busy railway in the Netherlands through many city centers, the experiments are divided into two sets. The first set of experiments focuses on the transit- and waiting times of trains on the Brabantroute, in order to give insights in the expected transit times on the Brabantroute. Set 2 focuses on synchromodality.

4.6.1 Set 1

The idea of the steering committee is to increase the utilization of the Brabantroute to 87.5 % (currently 75.0 %), such that it can handle more trains. A higher utilization makes a logistic system however more vulnerable to disruptions, only systems with a low variability can have a high utilization (Hopp & Spearman, 2008). Furthermore, it is plausible that the availability of the Brabantroute will decrease, since more trains could lead to more incidents due to a higher chance of failures and accidents. Set 1 therefore aims to examine the transit times of trains on the Brabantroute under different utilizations and disruptions. The design of the experiments of set 1 can be found below in Table 4.2. In total, 18 experiments are performed (6 different utilizations * 3 different availability's). Experiment 1 is the experiment with the planned utilization of the Brabantroute the coming years and the current availability.

Experiment Availability Brabantroute Utilization Brabantroute

1 94.17 % 87.5 % 2 94.17 % 90.0 % 3 94.17 % 85.0 % 4 94.17 % 82.5 % 5 94.17 % 80.0 % 6 94.17 % 75.0 % 7 97.00 % 87.5 % 8 97.00 % 90.0 % 9 97.00 % 85.0 % 10 97.00 % 82.5 % 11 97.00 % 80.0 % 12 97.00 % 75.0 % 13 92.00 % 87.5 % 14 92.00 % 90.0 % 15 92.00 % 85.0 % 16 92.00 % 82.5 % 17 92.00 % 80.0 % 18 92.00 % 75.0 %

(23)

23

4.6.2 Set 2

Set 2 of the experiments focuses on the actual comparison between a non-synchromodal solution and different synchromodal solutions. The three dimensions (time, route, modality) of synchromodality, introduced in section 2.3.4, are the main variables. Flexibility of route in this study could also be described as freedom of terminal, i.e. the container is allowed to be placed on another terminal. When there is no flexibility of route, a container is hence not allowed to enter another terminal (Scenario 1). The other variable is the number of trains per week that still can departure from Rotterdam to Duisburg. Due to the complexity of the disruptions of the Betuweroute and the uncertainty of priorities between different types of trains (container, wet bulk, hazardous materials) on the Betuweroute, it is uncertain how many Duisburg trains of EGS can still departure. Therefore, this variable is taken as a variable instead of a fixed parameter, like suggested by Robinson (2004:99). The experiments are performed for week two and three, in order to examine whether there will be differences in performance between weeks (even though the number of Duisburg trains is taken as a variable). These two weeks are chosen, since in these weeks the disruptions are most severe (see Table 3.1), respectively a complete week of only 40 trains per day and a week with a 0 trains in the weekend. The planned utilization and the availability of experiment1 (set1) are used as parameters for the Brabantroute. An overview of the different scenarios can be found in Table 4.3.

Scenario Modality free Route free Time freedom

1 Non-synchromodal alternative: only

rerouting via Venlo and then continue to Duisburg without stopping in Venlo

No No No 2 Yes No No 3 No Yes No 4 Yes Yes No 5 Yes No Yes 6 No Yes Yes

7 Full synchromodal Yes Yes Yes

Table 4.3: Scenarios set 2 experiments

(24)

24

In scenario 1, containers have no freedom of modality, route or time. This scenario could be viewed as the current situation with the addition of a train with final destination Duisburg that travels via the Brabantroute without stopping at the terminal in Venlo. Containers, ordered by customers, have a fixed modality and route and the due date is fixed. In scenario 2, containers may be transported on different modalities. Since containers that have a fast-approaching due date cannot be placed on a barge, a categorization is made. This categorization is made on the 'time till due date' (user-defined attribute), the time between the release date and the due date, and is based on discussions with planners. For e.g. containers with final destination Venlo the categorization can be found in Table 4.4. This categorization is applicable for all scenarios where freedom of modality is allowed.

Time till due date Possible modalities

< 18 hours Truck

> 18 hours and < 36 hours Truck, train

> 36 hours Truck, train, barge

Table 4.4: Categorization containers Venlo

In scenario 3 there is freedom of route. This implies that e.g. container with final destination Duisburg may be placed on a train to Venlo, unloaded at the terminal in Venlo and subsequently loaded on a truck to Duisburg. As stated earlier, a better description of freedom of route in this study is freedom of terminal. In scenario 4, 5 & 6, the due date of containers is stretched with 30 % compared to the previous scenarios. In reality, this could be achieved by e.g. setting the deadline of the booking date earlier.

4.7 Experimental settings

In order to obtain accurate output, the right experimental settings are a prerequisite. The simulation model examines a non-terminating process. This means that activities continue 24/7 and thus that the final settings of a day are the start settings of the next day. Because the model is a non-terminating process, a warm up period is needed in order to start the model in a realistic condition: when normal working conditions have been reached. The length of the warm up period is determined by using the Welch method (Robinson, 2004:146). The warm up period is set at two days. The total run length is 9.5 days (2 days warm up and 0.5 days to prevent end of horizon effect).

(25)

25

(26)

26

5. Results

In this section, the outcomes of the experiments will be presented. Firstly, the outcomes of set 1 will be presented and thereafter the results of set 2. For set 2, only the outcomes of the experiments with 5 Duisburg trains for week 2 will be presented, in order to create overview and make comparisons between the scenarios easier.

5.1 Outcomes set 1

The first set of experiments focuses on the waiting times of trains on the Brabantroute. The outcomes can be found in Table 5.1 below. In Figure 5.1, the average waiting time in minutes is plotted against the availability of the Brabantroute, for the five different utilizations. Besides the average waiting time, Figure 5.2 depicts the percentage of trains on the Brabantroute that is expected to have a waiting time higher than two hours. This number is more important than the average waiting time, since a train will be cancelled if the waiting time is too high and thus gives more insight than the average waiting time. Because common random numbers are used, pair wise paired T-tests are performed. Results indicate that the outcomes between different availability's and corresponding utilization are significant (α< 0.05). Furthermore, within every availability class (e.g. 94.17%) differences between different utilizations are also all found to be significant. A complete overview of the outcomes can be found in Appendix B.

Figure 5.1: Average waiting time trains 0 10 20 30 40 50 60 97.00% 94.17% 92.00% Wai tn g t im e ( m in u te s) Availability

Average waiting time Brabantroute

(27)

27

Experiment Availability Utilization Average waiting time (minutes) % trains waiting time > 2 hours 1 94.17 % 87.5 % 23.63 6.13 2 94.17 % 90.0 % 30.50 7.62 3 94.17 % 85.0 % 14.87 4.25 4 94.17 % 82.5 % 10.44 3.31 5 94.17 % 80.0 % 7.69 2.40 6 94.17 % 75.0 % 4.67 1.52 7 97.00 % 87.5 % 17.19 5.10 8 97.00 % 90.0 % 22.89 6.24 9 97.00 % 85.0 % 12.42 3.60 10 97.00 % 82.5 % 7.20 2.64 11 97.00 % 80.0 % 4.78 1.80 12 97.00 % 75.0 % 2.32 0.98 13 92.00 % 87.5 % 36.69 10.05 14 92.00 % 90.0 % 49.58 13.67 15 92.00 % 85.0 % 23.58 6.20 16 92.00 % 82.5 % 16.47 4.81 17 92.00 % 80.0 % 11.65 3.19 18 92.00 % 75.0 % 8.34 2.1

Table 5.1: Outcomes experiments Set 1

Figure 5.2: Percentage of trains with a waiting time longer than 2 hours 0.00% 2.00% 4.00% 6.00% 8.00% 10.00% 12.00% 14.00% 16.00% 97.00% 94.17% 92.00% % d e lay > 2 h o u rs Availability

% delay > 2 hours Brabantroute

(28)

28

As can be seen in the figures, the average waiting time and the percentage of trains with a waiting time larger than two hours both significantly increase when the availability of the Brabantroute decreases. Especially the percentage of trains with a waiting time larger than 2 hours is important. With the expected utilization of 87.5 % and the current availability of 91.17% (red line), 6.4% of the freight trains on the Brabantroute will have a delay longer than two hours. This number will further increase when the current availability will not be achieved.

5.2 Outcomes set 2

The outcomes of the experiments of set 2 will be presented in this section. The outcomes of all scenarios with five, four and three Duisburg trains are shown for week 2 (see respectively Table 5.2, Table 5.3 and Table 5.4). A full overview of the outcomes of the experiments of set 2 can be found in Appendix C and Appendix D.

Pair wise paired T-tests are performed to test for significance between the different scenarios within the same week with the same number of Duisburg trains. Results between scenario three and six are not significant, which also applies for the results between scenarios five and seven. Differences between all other scenarios are found to be significant (α<0.05). Furthermore, the differences between experiments with another number of Duisburg trains are small (i.e. the difference between experiment 1, 8 and 15). Finally, results between week 2 and 3 are insignificant (see Appendix C and D). There is only a small difference in on-time performance (week 3 somewhat lower). This is due to the fact that in week 3 no Duisburg trains departure in the weekend, such that containers waiting for the Duisburg train have to wait longer for the next train to departure.

(29)

29

Figure 5.3: Outcomes set 2: costs and on-time performance experiment 1-7 0 10 20 30 40 50 60 70 80 90 100 € - € 50,000 € 100,000 € 150,000 € 200,000 € 250,000 1 2 3 4 5 6 7 On -tim e p e rfor m an ce (% ) To tal co sts ( € ) Scenario

Results scenarios Set 2

Total costs (€) On-time

(30)

Scenario Exp. Modality flex. Route flex. Time flex. On-time (%) Number of trucks Total costs (€) Containers rerouted (%) Util Venlo train Util Duisburg train Util VD1 train Util RVD2 train Util Venlo barge Util Duisburg barge 1 1 No No No 68.8 356 _234,652 _3.22 _0.91 _0.92 _n/a _0.78 0.89 0.31 2 2 Yes No No 81.7 254 _213,593 _0.06 ₁ ₁ _n/a _0.01 0.93 0.39 3 3 No Yes No 73.2 167 _197,856 _4.47 _0.94 _0.92 _0.34 _0.79 0.89 0.31 4 4 Yes Yes No 81.8 241 _208,055 _0.08 ₁ ₁ _0.01 _0.01 0.93 0.39

5 5 Yes No Yes 91.6 83 _161,945 _0.00 ₁ ₁ _n/a ₀ 0.96 0.47

6 6 No Yes Yes 80.1 159 _196,698 _4.47 _0.94 _0.92 _0.38 _0.77 0.89 0.31

7 7 Yes Yes Yes 91.6 83 _161,945 _0.00 ₁ ₁ ₀ ₀ 0.96 0.47

Table 5.2: Outcomes experiments set 2 - 5 Duisburg trains

1

Venlo-Duisburg train

(31)

31

Scenario Exp. Modality flex. Route flex. Time flex. On-time (%) Number of trucks Total costs (€) Containers rerouted (%) Util Venlo train Util Duisburg train Util VD3 train Util RVD4 train Util Venlo barge Util Duisburg barge 1 8 No No No 66.2 356 _229,652 _3.78 _0.91 _0.94 _n/a _0.83 0.89 0.31 2 9 Yes No No 81.3 269 _213,847 _0.09 ₁ ₁ _n/a _0.01 0.93 0.42 3 10 No Yes No 70.9 167 _192,856 _4.93 _0.94 _0.94 _0.37 _0.85 0.89 0.31 4 11 Yes Yes No 81.3 244 _204,452 _0.08 ₁ ₁ _0.01 _0.01 0.93 0.42

5 12 Yes No Yes 91.3 85 _157,545 _0.00 ₁ ₁ _n/a ₀ 0.96 0.49

6 13 No Yes Yes 77.7 159 _190,698 _4.93 _0.94 _0.94 _0.41 _0.85 0.89 0.31

7 14 Yes Yes Yes 91.3 85 _157,545 _0.00 ₁ ₁ ₀ ₀ 0.96 0.49

3

(32)

32

Scenario Exp. Modality flex. Route flex. Time flex. On-time (%) Number of trucks Total costs (€) Containers rerouted (%) Util Venlo train Util Duisburg train Util VD5 train Util RVD6 train Util Venlo barge Util Duisburg barge 1 15 No No No 64.3 356 _224,652 _4.03 _0.91 _0.97 _n/a _0.88 0.89 0.31 2 16 Yes No No 81.2 284 _213,569 _0.11 ₁ ₁ _n/a _0.02 0.93 0.45 3 17 No Yes No 68.6 167 _187,586 _5.21 _0.94 _0.97 _0.41 _0.91 0.89 0.31 4 18 Yes Yes No 80.9 248 _201,867 _0.10 ₁ ₁ _0.02 _0.02 0.93 0.45

5 19 Yes No Yes 90.8 86 _152,843 _0.00 ₁ ₁ _n/a ₀ 0.96 0.53

6 20 No Yes Yes 74.8 159 _185,698 _5.21 _0.94 _0.97 _0.44 _0.91 0.89 0.31

7 21 Yes Yes Yes 90.8 86 _152,843 _0.00 ₁ ₁ ₀ ₀ 0.96 0.53

5

(33)

As can be seen in Figure 5.3, the non-synchromodal alternative (scenario 1) has the highest costs and lowest on-time performance. Scenario 2, freedom of modality, performs significantly better. Scenario 3, freedom of route, has lower total costs than scenario 2, but the on-time performance is worse. Scenario 4 performs slightly better than scenario 2 in terms of costs. When the due date is stretched (scenario 5, 6 and 7), performance increases significantly. The results show that when there is a stretched due date and modality is flexible, route flexibility makes no difference anymore: scenario 5 and 7 have the same performance.

(34)

34

modality of container is fixed. Finally, no difference is found between scenario 5 and 7. The decrease of Duisburg trains can be absorbed by the flexibility in modality when there is a stretched due date (because more containers are allowed on all three modalities).

(35)

35

6. Discussion and limitations

Set 1 of the performed experiments shows that, with the same availability and a higher utilization than currently on the Brabantroute, 6.13 % of all trains that will departure on this route are expected to have a delay of more than two hours. This is in line with the core principles of operations management. Hopp and Spearman (2008) state that a logistic system is very vulnerable for interruptions when the utilization exceeds 80 %. It is however questionable if the same availability will be achieved with a higher utilization. It is likely that more incidents will happen on the railway when more trains travel on it, which subsequently will decrease the current availability. More trains can lead to more failures of equipment and the chance on accidents will be higher as well. A lower availability could therefore be more realistic, resulting in even more trains that will have a delay longer than two hours.

(36)

36

Focusing on flexibility in modality is the most important managerial insight, together with flexibility in time. Managers should focus on these two dimensions in order to achieve the benefits of synchromodal transportation. The main barriers to achieve flexibility on these dimensions are the customers, who mainly determine the booking in the current situation. Therefore, important benefits like improved reliability should be clearly shown to them. This study can help to show these benefits. Complete synchromodal transportation on all three dimensions will be difficult to achieve due to e.g. tight due dates and powerful customers. However, if only a part of the containers have flexibility in modality, significant improvements can be made. Managers should thus target at flexibility in modality and time by designing a product portfolio with different tariffs. Combining a differentiated portfolio with a clear reasoning of the benefits for the customers will lead to more flexibility and hence a better performance.

(37)

37

7. Conclusion

(38)

38

References

Baker, K. R. 1984. Sequencing rules and due-date assignments in a job shop. Management science, 30(9), 1093-1104.

Boysen, N., Fliedner, M., Jaehn, F., & Pesch, E. 2013. A Survey on Container Processing in Railway Yards. Transportation Science, 47(3), 312-329.

Commission of the European Communities. 1997. Intermodality and intermodal freight transport in the European Union. Communication 243, Brussels.

Craig, A. J., Blanco, E. E., Sheffi, Y. 2013. Estimating the CO2 intensity of intermodal freight transportation. Transportation Research Part D-Transport and Environment, 22, 49-53. Das, S. K., &Nagendra, P. 1997. Selection of routes in a flexible manufacturing facility. International journal of production economics, 48(3), 237-247.

European Gateway Services.2014. http://www.europeangatewayservices.com/#view=map . Website accessed at Wednesday October 14, 2014.

Europe Container Terminals. 2013. Key figures edition 2013. Eurostat. October 2012. Freight Transport Statistics.

Eurostat. November 2013. Freight Transport Statistics – Modal Split Eurostat. February 2014. Maritime ports freight and passenger statistics.

Hopp, W.J., Spearman, M.L. 2008. Factory Physics, International Edition – Third Edition. McGraw-Hill Education – Europe.

Keskinocak, P., Tayur, S. 2004. Due Date Management Policies. Handbook of Quantitative Supply Chain Analysis. International Series in Operations Research & Management Science,77, 485-554.

Keyrail. 2015. http://www.keyrail.nl/over_keyrail/facts_figures. Website accessed at Tuesday January 13, 2015.

(39)

39

Koste, L. L., &Malhotra, M. K. 1999. A theoretical framework for analyzing the dimensions of manufacturing flexibility. Journal of operations management,18(1), 75-93.

Langen, P. W. de. 2004. The Performance of Seaport Clusters; A Framework to Analyze Cluster Performance and an Application to the Seaport Clusters of Durban, Rotterdam and the Lower Mississippi. (No. ERIM PhD Series; EPS-2004-034-LIS). Erasmus Research Institute of Management (ERIM).

Lattila, L., Henttu, V., &Hilmola, O. 2013. Hinterland operations of sea ports do matter: Dry port usage effects on transportation costs and CO2 emissions. Transportation Research Part E-Logistics and Transportation Review, 55, 23-42.

Liao, C., Tseng, P., & Lu, C. 2009. Comparing carbon dioxide emissions of trucking and intermodal container transport in Taiwan. Transportation Research Part D-Transport and Environment, 14(7), 493-496

Lucassen, I.M.P.J., Dogger, T. 2012. TNO Report: Synchromodality pilot study.

Macharis, C., Bontekoning, Y.M. 2004. Opportunities for OR in intermodal freight transport research: A review. European Journal of Operational Research, 153: 400–416.

Port of Rotterdam. 2013. Havenvisie 2030: Voortgangsrapportage 2013.

http://www.portofrotterdam.com/nl/Over-de-haven/haven-rotterdam/havenvisie2030/Documents/Voortgangsrapportage2013-Havenvisie2030.pdf

Reis, V. 2014. Analysis of mode choice variables in short-distance intermodal freight transport using an agent-based model. Transportation Research Part A: Policy & Practice, 61, 100-120.

Riessen, B. van. 2013. Planning of Hinterland transportation in the EGS network. Master thesis, Delft University of Technology and Erasmus University Rotterdam.

(40)

40

Riessen, B. van, Negenborn, R.R., Dekker, R., Lodewijks, G. 2013b. Analysis of impact and relevance of transit disturbances in an intermodal container network. 3rd International Conference on Logistics and Maritime Systems (LOGMS 2013), Singapore.

Riessen, B. van, Negenborn, R.R., Dekker, R. 2014. Poster Synchromodal container transportation. SmartPort poster session, October 2014, Rotterdam.

Rijksoverheid. 2014.Website accessed at Thursday October 16, 2014.

http://www.rijksoverheid.nl/documenten-en-publicaties/brieven/2014/10/14/derde-spoor-duitsland-advies-stuurgroep-aan-staatssecretaris.html

Robinson, R. 2002. Ports as elements in value-driven supply chain systems: the new paradigm. Maritime Policy and Management 25, 21–40.

Robinson, S. 2004. Simulation the practice of model development and use. Chichester, Eng.; Hoboken, N.J.: Wiley.

Roso, V., Woxenius, J., Lumsden, K. 2009. The dry port concept: connecting container seaports with the hinterland. Journal of Transport Geography, 17(5), 338-345.

SteadieSeifi, M., Dellaert, N. P., Nuijten, W., Van Woensel, T., Raoufi, R. 2014. Multimodal freight transportation planning: A literature review. European Journal of Operational Research, 233(1), 1-15.

Tongzon, J. L. 2009. Port choice and freight forwarders. Transportation Research Part E: Logistics and Transportation Review, 45(1), 186-195.

Topteam Logistiek. 2011. Partituur naar de top. Adviesrapport Topteam Logistiek.

Verweij, K. 2011. Synchronic modalities – Critical success factors. In P. J. van der Sterre (Ed.), Logistics yearbook edition 2011. Rotterdam (pp. 75–88).

Vetevood, S. 2008. Intermodal Freight Transportation – Requirements for Tactical Evaluation Tool. Master thesis, Technical University of Denmark.

(41)

41

Zhang, R., Yun, W. Y., Moon, I. 2009. A reactive tabu search algorithm for the multi-depot container truck transportation problem. Transportation Research Part E: Logistics and

(42)

Appendix A–Experimental design experiments set 2

Week 2 Experiment Number of Duisburg trains Modality flexibility Route flexibility Time

flexibility Week 3 Experiment

Number of Duisburg trains Modality flexibility Route flexibility Time flexibility 1 5 No No No 22 5 No No No 2 5 Yes No No 23 5 Yes No No 3 5 No Yes No 24 5 No Yes No

4 5 Yes Yes No 25 5 Yes Yes No

5 5 Yes No Yes 26 5 Yes No Yes

6 5 No Yes Yes 27 5 No Yes Yes

7 5 Yes Yes Yes 28 5 Yes Yes Yes

Week 2 8 4 No No No Week 3 29 4 No No No

9 4 Yes No No 30 4 Yes No No

10 4 No Yes No 31 4 No Yes No

14 4 Yes Yes Yes 35 4 Yes Yes Yes

Week 2 15 3 No No No Week 3 36 3 No No No

16 3 Yes No No 37 3 Yes No No

17 3 No Yes No 38 3 No Yes No

(43)

43

Appendix B - Outcomes experiments set 1

Availability Utilization Experiment

(44)

44

Appendix C - Outcomes experiments set 2 week 2

7 7 SG = South-Germany Week 2 5 Duisburg

trains Experiment Flexibility

Number

of trucks Total cost

On-time performance (%) Util Dui Trains Util Ven Trains Util Venlo barge Util Duisburg barge Util RVD Trains Util Enns trains Util Vienna trains Util SG Trains Util VD Trains % Reroute Waiting time containers (hours) Scenario 1 1 - 356 € 234,652 68.8 0.92 0.91 0.89 0.31 0.78 0.91 0.58 0.32 n/a 3.22 15.69 Scenario 2 2 Modality 254 € 213,593 81.7 1 1 0.93 0.39 0.01 0.91 0.58 0.32 n/a 0.06 12.25 Scenario 3 3 Route 167 € 197,856 73.2 0.92 0.94 0.89 0.31 0.79 0.91 0.58 0.32 0.34 4.47 13.38 Scenario 4 4 Modality + route 241 € 208,055 81.8 1 1 0.93 0.39 0.01 0.91 0.58 0.32 0.01 0.08 12.13 Scenario 5 5 Modality + time 83 € 161,945 91.6 1 1 0.96 0.47 0 0.91 0.58 0.32 n/a 0 12.58 Scenario 6 6 Route + time 159 € 195,698 80.1 0.92 0.94 0.89 0.31 0.77 0.91 0.58 0.32 0.38 4.47 13.21 Scenario 7 7 Modality + route + time 83 € 161,945 91.6 1 1 0.96 0.47 0 0.91 0.58 0.32 0 0 12.58

Week 2 4 Duisburg

Number

(45)

45

Number

(46)

46

Appendix D - Outcomes experiments set 2 week 3

Number

On-time performance (%) Util Dui Trains Util Ven Trains Util Venlo barge Util Duisburg barge Util RVD Trains Util Enns trains Util Vienna trains Util SG Trains Util VD Trains % Reroute Waiting time containers (hours) Scenario 1 22 - 368 € 239,975 68.5 0.91 0.91 0.89 0.31 0.80 0.91 0.58 0.32 n/a 3.31 15.73 Scenario 2 23 Modality 254 € 213,593 81.6 1 1 0.93 0.39 0.01 0.91 0.58 0.32 n/a 0.06 12.35 Scenario 3 24 Route 167 € 197,856 73.1 0.92 0.94 0.89 0.31 0.81 0.91 0.58 0.32 0.35 4.55 13.45 Scenario 4 25 Modality + route 241 € 208,055 81.7 1 1 0.93 0.39 0.01 0.91 0.58 0.32 0.01 0.08 12.17 Scenario 5 26 Modality + time 83 € 161,945 91.6 1 1 0.96 0.47 0 0.91 0.58 0.32 n/a 0 12.61 Scenario 6 27 Route + time 159 € 195,698 80.0 0.92 0.94 0.89 0.31 0.77 0.91 0.58 0.32 0.39 4.55 13.24 Scenario 7 28 Modality + route + time 83 € 161,945 91.5 1 1 0.96 0.47 0 0.91 0.58 0.32 0 0 12.61

Number

(47)

47

Number

A simulation to examine a synchromodal hinterland transportation network

Master’s Thesis