Operational Aspects in Cross-‐Docking Networks: A
Literature Review and Case Study
Master’s Thesis Operations & Supply Chain Management
University of Groningen, Faculty of Economics and Business
Franciscus Jozef Egidius Rutten
December 20
th, 2013
Contact Details:
Koestraat 18,
8011 NK Zwolle
e-‐mail: f.j.e.rutten@gmail.
comSupervisors & Assessors:
Prof. Dr. Ir. J.C. Wortmann (Supervisor, First assessor)
Dr. N.B. Szirbik (Co-‐assessor)
Abstract
Less Than Truckload (LTL) cross-‐docks are faced with complex operations and are therefore subjected to many interrelated management decisions. In order to cope with this complexity, academic literature proposes mathematical models aimed at supporting cross-‐dock decisions. However, few of these models are encountered in LTL cross-‐docking practice. Through a literature review and an in-‐depth case study at an LTL cross-‐dock, this thesis explores why the decision support models from academic literature are not widespread among LTL cross-‐docks. Differences in stakeholder, network design, and cross-‐dock characteristics have been found to be among the core reasons. The fact that decision models are typically address a single cross-‐dock decision problem aspect further reduces applicability of decision support models in industry. For LTL cross-‐dock managers, this thesis provides guidance into the typical objectives, decisions problem aspects, and assumption considered in academic literature and formulates recommendations to enhance the application of the academic decision models in industry. For academic developers of cross-‐dock decision models, this thesis presents detailed insight into a typical LTL cross-‐dock organization and provides suggestions for future innovations with industrial relevance.
Keywords: Transportation; LTL cross-‐dock; decision support model; literature review; single case
Acknowledgements
To begin this paper, and before continuing with the actual content of this Master Thesis, I would like to take this opportunity to thank everyone who has contributed to the successful completion of this research. First of all, I would like to thank my supervisor Paul Buijs for giving me a lot of support and for the interesting suggestions and feedback he provided throughout my research. A similar gratitude is also in place for my second supervisor, Prof. Dr. Ir. Hans Wortmann and my co-‐assessor Dr. Nick Szirbik, who provided me with in-‐depth feedback and valuable insights during my Master Thesis project. Thirdly, I would like to thank dr. Manda Broekhuis for her continuing support over the past few years.
Also, I would like to show my appreciation to all colleagues who provided me with information during my 4 months internship at the cross dock under study; the material handlers, the cross dock supervisors and especially the cross dock manager, who gave me every day an hour ride to-‐ and from the cross dock for which I am very grateful.
Finally, I would like to thank my friends and family who supported me during the, somewhat long, Master Thesis project. It was their support that encouraged me to eventually finish this thesis. A special thanks goes to my girlfriend and her everlasting faith in me.
Thank you all very much.
Table of contents
Abstract ... 2
Acknowledgements ... 3
Table of contents ... 4
Glossary ... 6
1. Introduction ... 7
2. Background ... 10
2.1 Objectives in cross-‐dock scheduling ... 10
2.2 Decision problem aspects ... 13
2.2.1 Dock door specification and dock door assignment ... 14
2.2.2 Sequencing ... 15
2.2.3 Inner transport ... 16
2.3 Characteristics and assumptions ... 18
3. Methodology ... 20
3.1 Research questions ... 20
3.2 Scope of the research ... 20
3.3 Methods ... 21
3.3.1 Structured literature review ... 21
3.3.2 Exploratory case study ... 22
3.3.3 Analytical generalization ... 23
4. Empirical findings ... 24
4.1 Characteristics of the LTL case company ... 24
4.1.1 Shipment characteristics ... 24
4.1.2 Goods flow characteristics ... 25
4.1.3 Trailer and dock door characteristics ... 26
4.1.4 Operation schedule characteristics ... 27
4.2 Process ... 29
4.2.1 The inner cross-‐dock process ... 29
4.2.2 Dock door assignment ... 31
4.2.3 Shunting ... 32
4.2.4 Key Performance Indicators ... 32
4.3 Generalization of LTL cross-‐docks ... 33
4.4 Comparison of cross-‐docks in literature and the case company ... 34
5. Propositions and recommendations ... 38
5.1 Propositions ... 38
5.1.1 Stakeholders ... 38
5.1.2 Network design ... 39
5.1.3 Cross-‐dock characteristics ... 41
5.2 Recommendations for academic research ... 43
5.3 Recommendations for the case company ... 44
6. Conclusion ... 47
6.1 Main research findings ... 47
6.2 Limitations and further research ... 48
References ... 49
Appendix A: Classification of cross-‐docking research ... 53
Appendix B: Informal discussions and observations ... 54
Appendix C: Shipments handled by the case company ... 56
Appendix D: Shipments handled by the case cross-‐dock ... 57
Glossary
Domestic trucks: Trucks serving domestic destinations and are typically performing delivery and collection runs
Due date: The latest moment in time a truck can leave the cross-‐dock to reach its destination(s) on time
Inbound dock door: Dock door used for unloading shipments from inbound trucks
Inbound truck: Incoming truck with shipments which have to be unloaded at the cross-‐dock
International trucks: Trucks serving international destinations and are typically shuttling between two cross-‐docks
Loading: Load shipments into their destined outbound truck
Material handlers: Employees employed at the cross-‐dock performing unloading, moving and loading activities
Order picking: Retrieve shipments from the storage area, and possible repacking activities
Outbound dock door: Dock door used for loading shipments in outbound trucks
Outbound truck: Outgoing truck which have to be loaded with shipments for delivering purposes
Sorting (moving): Move shipments to a storage area
Storing: Keep shipments in storage
Unloading: Unloading shipments from inbound trucks
1. Introduction
In recent years, deregulation under European Union (EU) law has led to increasing competition in freight service markets in the EU. Therefore, transportation companies must achieve high performance levels in terms of economic efficiency and quality of service. The former because a transportation firm must make profit in an open, competitive and mainly cost driven market. The latter, because transportation services must conform the high standards that are expected by current management strategies, such as small or no inventory across the supply chain along with a just-‐in-‐time production strategy. The introduction and use of the cross-‐docking strategy is an example of the adaptations made by transportation companies to improve profit and service levels in reaction to the increased standards (Saddle Creek Logistics, 2008; 2011). Cross-‐docking can realize transport efficiencies and reduced material handling and storage costs by eliminating the storage and order picking activities from the main warehouse operation (Apte and Viswanathan, 2000). Cross-‐docking is used in dedicated retail supply chains as well as in less-‐than-‐truckload (LTL) transportation environments.
of 24 hours. In a warehouse the stored products are known and the moment at which the customer order arrives is unknown; while in a cross-‐dock environment the product arrival is unknown, but the departure times are known (Vis and Roodbergen, 2008). Accordingly, supply chains based on the cross-‐dock system are seamlessly coupled from supplier to consumer (Vogt, 2010). In literature, global companies such as Wal Mart (Stalk et al., 1992), Eastman Kodak Co. (Cook et al, 2005), Toyota (Witt, 1998) and UPS (Forger, 1995) have reported successful implementations of cross-‐docking systems. The reader is referred to Van Belle et al. (2012) and Buijs et al. (2013) for an overview of cross-‐docking literature.
The cross-‐dock operation is subjected to many management decisions. The planned and actual arrival time of inbound trucks, due dates of outbound trucks, number of dock doors, number of forklifts, the size and hand ability of shipments, and the destination of individual shipments are just a few examples. Due to the high number of decisions it is complex to efficiently operate a cross-‐ dock. Therefore, in most cross-‐dock operations some form of decision support is indispensable. While the cross-‐dock system has been increasingly applied in supply chains, scientific interest in the subject grew with it. Over the last two decades a fast growing number of decision support models is described in literature. However, few of these models are encountered in LTL cross-‐docking practice (Buijs and Vis, 2013). This is conflicting with the general view that, due to the complexity of operations, decision support is needed to efficiently operate a cross-‐dock. In this thesis, we are studying this contradiction by comparing a single in-‐depth case study at an LTL carrier and academic perspectives on mathematical decision support for cross-‐docking operations in general. To the best of our knowledge, no other empirical studies have been performed on this subject thus far. The primary objective of this thesis is twofold.
1) First, it aims to explore which aspects of existing mathematical decision models can be used in LTL cross-‐dock practice and which changes to physical operations or management approaches are required to make it work.
2) Second, it aims to explore which fundamental changes are needed in future decision modeling to enhance applicability in the LTL cross-‐dock industry.
The single in-‐depth case study is conducted at an LTL cross-‐dock in the Netherlands. Although a single case limits generalizability in addressing the research objective, it allows for a detailed exploration. Accordingly, the empirical findings will be generalized to certain extend by means of analytical generalization techniques. The cross-‐dock under study is characterized by:
- When a shipment is unloaded at the cross-‐dock, the designated outbound truck is in general not known yet
- On average, a typical dock door handles 3 trailers a night
As will be described in the remainder of this thesis, there are some fundamental differences between the characteristics of the cross-‐dock under study and those of cross-‐dock settings often described in literature. Therefore, the aim of this thesis is to answer the following research question:
“To what extend does synergy exist between academic cross-‐docking literature and the operations of the LTL cross-‐dock under study, and what opportunities for improvement exist?”
The main contribution of this thesis is extending academic literature on cross-‐dock decision modeling by investigating if– and how– existing decision models are applicable into LTL cross-‐ docking practice. As described, this topic has received little attention in the literature. The outcomes of this research may give insights to LTL cross-‐docks for implementing cross-‐dock decision modeling into their organization. Besides the relevance for cross-‐docking practice, the insights provided by this research could prove highly relevant for academic literature as well. By offering insight into the operations of LTL cross-‐docking practice, researchers in the field of cross-‐docking could make a more adequate assessment of the chosen constraints and assumptions in their decision models. Moreover, this study provides an understanding of how cross-‐dock operations in practice are organized. This knowledge can be applied by academic researchers involved in the research of cross-‐ dock decision modeling to develop decision models which are more applicable in cross-‐docking practice.
The remainder of this thesis is structured as follows. In section 2 the theoretical background of this study will be outlined. Subsequently, chapter 3 discusses the methodology. This chapter also elaborates on the single in-‐depth case study which was conducted for this research. Chapter 4 presents the empirical data gathered during the case study, followed by propositions and recommendations in Chapter 5. The conclusion and discussion of this study as well as directions for future research are summarized in Chapter 6.
2. Background
Cross-‐docking is a comparatively new logistics strategy in practice and also the body of academic literature written about this topic is not yet mature. In fact, the first dedicated research on the short-‐ term truck scheduling problem was published in 2005 (McWilliams, 2005). Because of differences in organizational and technical implementations there is a large variety of possible planning problems in real-‐world settings. Therefore existing literature considers different objectives, decision problems, constraints and assumptions. This section outlines the main decision support models described in academic literature. First, the main objectives in cross-‐dock scheduling are elaborated upon in order to provide a better insight into the outlines of current academic literature. Thereafter, different decision problem aspects are explored, including their constraints and assumptions. This, to give an overview of the specific aspects which may be used to achieve a certain objective. Finally, the most recurring cross-‐dock characteristics and assumptions in literature are discussed.
Due to the readability of this thesis, only the most needful references are added to the textual elements. When is spoken of “typical cross docks in literature”, the reader is referred to appendix A which classifies the assessed academic cross docking literature.
2.1 Objectives in cross-‐dock scheduling
In this paragraph, the main objectives in cross-‐dock scheduling are elaborated upon in order to provide a better insight into the outlines of current academic literature. Cross-‐dock scheduling decides on the sequence and assignment of incoming and outgoing trucks at the dock doors of the cross-‐docking terminal, subject to the availability of cross-‐docking resources including dock doors and material handling systems.
scheduling literature which are all focusing on minimizing operational costs of the cross-‐dock facility. These objectives are presented in table 1.All four objectives have a cost related objective function.
Objective Purpose Goal
Minimizing the makespan
Minimizing the time interval between unloading the first shipment and loading the last shipment in a planning horizon
Shortening the makespan at a fixed workforce capacity rate implies increased productivity and thereby decreased operational costs.
Minimizing inner
travel distance Minimizing the total distance travelled by forklifts moving shipments from inbound to outbound dock doors during the entire makespan
Reduction of the average moving distance of shipments through the cross-‐dock and thereby reducing the average handling time of individual shipments. This will improve the average productivity of the material handlers which in turn will have a positive influence at the workforce of the company.
Minimizing earliness and tardiness
Minimizing the penalty for
trucks leaving to late or to early Minimizing penalty costs
Minimizing
inventory Minimizing the total amount of temporary inventory stored at the cross-‐dock
Reduction of the chance of delayed shipments, because shipments are directly moved to the destined outbound trailer, instead of putting the shipments in temporary storage.
Minimizing inventory will led to a minimization of the number of shipments per unit time in temporary storage. The reduced stock size will reduce the risk of congestion of forklifts inside the terminal. Cross-‐docks with limited storage facilities may use this objective to schedule trucks in a way that the maximum storage capacity is not exceeded instead of expanding the temporary storage facilities.
Table 1: The main objectives in academic cross-‐docking literature
The makespan and the inner travel distance objective, which is as discussed in table 1 in close relation to the workforce component, are of great influence at the throughput rate of the cross-‐dock. In literature, the throughput rate is generally viewed as the rate at which shipments are moved through the -‐dock at a given workforce within a given time unit. Hence, together with the
total freight volume (i.e., the quantity and volume of shipments that have to be handled at the cross-‐
dock within a given planning horizon) the makespan and workforce determine the throughput rate of a cross-‐dock. Improving any of these three effects has a positive outcome on the operational costs of the cross-‐dock: a lower workforce results in lower total salaries costs and the need for less material handling equipment, a decrease in makespan is shortening the work shifts and less other operational costs and a higher volume results in more turnover at the same operational costs. A combination of the three factors is also possible. Hence, this implies that the throughput rate has a direct influence at the operational costs of a cross-‐dock organization.
Figure 1: the influence of throughput rate at the operational costs
Cross-‐docks facilitate consolidation at the lowest possible costs and therefore fulfill a dedicated role in the supply chain. However, the volume of shipments that have to be consolidated are determined by the suppliers’ demand. Since all shipments typically have to be consolidated and loaded in outbound trucks within 24 hours after arriving, cross-‐docks do not have influence on product flow of their clients and thereby on the volume of shipments which have to be transferred during a transfer operation. That is likely the reason why there is virtually no known cross-‐dock literature aimed at improving the volume of shipments. However cross-‐dock literature discusses decision problems affecting the makespan and the workforce of cross-‐docks during a transfer operation.
Based on this, it can be concluded that the main objectives in cross-‐dock scheduling are the minimization of the makespan and inner travel distance. Together with the total freight volume they determine the throughput rate of the terminal which in turn is of influence at the operational costs.
2.2 Decision problem aspects
Decision problem aspects are important aspects in a cross-‐dock facility because they enable cross-‐ docks to achieve certain objectives. Boysen and Fliedner(2010) and Bartholdi and Gue (2000) for example describe multiple decision problems which are affecting the freight flow and therefore the workload within a cross-‐dock facility. In turn, the freight flow has an effect on the makespan, the required workforce and the throughput rate of the cross-‐dock. The decision problem aspects are shown in table 2.
Decision problem aspect Determines Influence
Layout of the terminal The number of dock doors and the shape of the terminal (e.g., I, T or X-‐shaped)
The layout indirectly determines the location of the doors and the travel distances between them. The shape of the terminal also affects the rate of congestion: narrow docks and pathways tend to be more congested because workers have less room to maneuver
Freight mix The composition and ratio of different shipment dimensions and characteristics
the average handling time of shipments. Oversized shipments will need more unloading, moving and loading time than a typical euro pallet
Material
handling equipment The composition of the handling equipment Influenced by the freight mix, taking the costs of the equipment and their disadvantages into account
Door specification Which door is used for inbound and which door is used for outbound activities
The cumulative distance of the shipments between the inbound dock and the destined outbound docks
Dock door assignment Which inbound truck has to unload at which unload dock, and which outbound truck has to load at which load dock
The cumulative distance of the shipments between the inbound dock and the destined outbound docks
Truck sequencing The order in which trucks dock
at their assigned dock door Synchronization of inbound and outbound trucks
Inner transport in which order shipments inside the terminal have to be moved
Departure times of outbound trucks Table 2: Decision problem aspects
dock door specification and dock door assignment are combined since they overlap slightly and serve the same objective.
2.2.1 Dock door specification and dock door assignment
As mention above, Boysen and Fliedner (2010) mention that dock door specification and dock door assignment are important operational decision problems for influencing the travel distance of shipments inside the cross-‐dock, and are of direct influence at the workforce component. Inbound trailers are typically carrying multiple shipments with different destinations that have to be moved through the cross-‐dock to different outbound trailers. With the knowledge of all destinations of incoming shipments the cross-‐dock is able to create an outbound truck schedule before the transfer operation starts. It is generally assumed in academic literature that dock doors are dedicated for inbound or outbound activities and that dock doors are identical, which implies that all trucks can dock at all dock doors. Although most cross-‐docking literature assign multiple destinations to outbound doors, some papers typically assign outbound doors exclusively to one destination. This fixed assignment eases the allocation of shipments to trucks, since employees can “learn” the topology of the terminal. Fixed assignments seem especially suited for steady commodity flows with a reliable distribution among inbound and outbound destinations (Apte and Viswanathan, 2000). However, a fixed assignment of doors to destinations restricts the degrees of freedom for short-‐term truck scheduling, because peak loads for single destinations cannot be absorbed by additional dock doors.
The number of dock doors and their placement along the perimeter of the terminal are known. Consequently, the distance between any pair of doors is given, so that the transfer distance between those doors can be anticipated accurately. Hence, the inbound and outbound truck assignment decision influences the inner travel distance of all shipments of an inbound load. The objective of de dock door assignment decision is to minimize the total inner travel distance of shipments between inbound and outbound dock doors. The shorter the distance from inbound to outbound dock the sooner the forklift can pick up another shipment from the inbound dock and thereby reducing not only the average travel distance, but also the average inner transportation time of shipments in the dock. Because of the decrease in average transportation time forklifts can move more shipments per time unit which in turn will increase the throughput rate of the cross-‐dock facility.
2.2.2 Sequencing
Sequencing is, as mentioned above influencing the makespan of the cross-‐dock (Boysen and Fliedner, 2010). The basic premise of truck sequencing algorithms is that the sequence in which available trailers are served is most important. The premise translates in the assumption that the number of trucks is greater than the number of dock doors. In solving the dock door assignment problem, most academic literature assumes that the inter-‐arrival time of the set of trailers at a single dock door is large enough to consider them as one. On the contrary, truck sequencing problems deal with the situation where the sequence in which trailers are served is important. Taking the above into consideration, deciding upon the sequence in which trucks are served is based on the premise that synchronization of inbound and outbound trucks is a good proxy for cross-‐dock performance. Where dock door assignment is focusing on minimizing the average distance between inbound and outbound dock doors and thereby minimizing the workforce component of the throughput rate, the truck scheduling problems are all focusing on improving the time-‐related performance indicator which is expressed in the makespan component of the throughput rate. However, current truck sequencing algorithms in academic literature (e.g. Larbi et al, 2011;Briskorn et al, 2010) are only applicable when all inbound and outbound trailers are available at time zero or the arrival sequence of the inbound trailers can be fully determined by the cross-‐dock.
Cross-‐docks are receiving inbound shipments, consolidate the shipments based on their destination and transport them to their designated outbound truck. Since there are in general not enough available doors to dock all inbound and outbound trailers at the same time, the sequence in which the inbound and outbound trucks dock is of great influence on the transshipment process. The main contribution of truck sequencing is the synchronizing of inbound to outbound shipment flows and is typically aiming at a minimization of the makespan of the transfer operation. However truck sequencing can also be used to minimize the inventory-‐ and earliness and tardiness objectives.
Minimization of inventory
shipments and thereby the total transshipment time. Congestion at the cross-‐dock can be minimized by effective truck sequencing. For example, by aggregate demand for each outbound door, inbound trucks can be sequenced in such a way that inbound shipments, and thereby the workload, are more or less proportionally distributed over the outbound docks, minimizing the average inventory levels and thereby floor space congestion (Wang and Regan, 2008). Hence, truck sequencing enables the cross-‐dock to become as close as possible to the situation in which the inbound truck and its destined outbound trucks are simultaneously docked, thereby minimizing the inventory and the additional negative aspects.
Minimizing earliness and tardiness
Satisfying service level requirements is one of the key performance indicators of Just-‐In-‐Time supply chains. In general, the most important aspect in this context is punctual running of outbound trucks delivering shipments to clients. For that reason Boysen en Fliedner (2010), and Boysen et al. (2013) describe situations in which outbound trucks have due dates which have to be met. Both leaving too early (earliness) and leaving too late (tardiness) have to be avoided, whereby in most cases tardiness outweighs earliness. Earliness and tardiness can direct and indirect cause additional costs. Direct costs can be caused by penalty costs and indirect costs by a reduction in goodwill. To minimize the earliness and tardiness in truck scheduling, the basic objective is to specify the sequences of inbound trucks in a way that the penalties for earliness and tardiness of outbound trucks are minimized.
It can be concluded that sequencing is mainly influencing the makespan of the cross-‐dock. However, the makespan also influences the minimization of inventory-‐ and earliness and tardiness objective.
2.2.3 Inner transport
assumed that outbound shipments have the same destination. This gives certain degrees of freedom during loading operations and inner transport because all shipments can directly be loaded in docked outbound trailers.
Academic literature on inner transport assumes that during a transfer operation there might be delays in shipment handling times due to the unavailability of employees and forklifts (e.g. Shakeri, 2012). If no forklift is available, a delay is incurred to the pallet loading time till the required resource is available. The amount of delays depends on the total number of forklifts operating in the cross-‐dock, how they are scheduled and on the due dates of the outbound trucks. By knowing the due dates of the outbound trucks, shipment transportation order can be composed which determines the moving order of the individual inbound shipments. When two shipments are in the staging area and are both ready to move, literature generally assumes that the shipment with the highest priority will be moved first. The objective is typically to complete each outbound truck exactly at its due date and thereby minimizing earliness and tardiness.
Because of narrow pathways in the cross-‐dock interference between forklifts may occur causing congestion between delivering forklifts and passing forklifts(Bartholdi and Gue, 2000). Narrow pathways may be enhanced by relatively large intermediate storage since this floor space cannot be used for moving shipments. Also the space to maneuver decreases, which mainly has effect on forklifts trying to pass a stack door where another forklift is unloading a trailer. Forklift scheduling policies may take forklift interference into account with the objective of minimizing the congestion. Minimizing congestion will achieve less downtime on forklifts, improve their productivity and thereby the makespan and the throughput rate of the cross-‐dock.
Figure 2: figure 1 extended with the influence of the decision problem aspects at the main cross-‐dock
scheduling objectives
This paragraph explored different decision problem aspects, including their constraints and assumptions. This, to give an overview of the specific aspects which may be used to achieve a certain objective.
2.3 Characteristics and assumptions
Finally, the most recurring cross-‐dock characteristics and assumptions in literature are discussed. To formulate the truck scheduling problem, most academic literature are sketching some operational characteristics and assumptions concerning their cross-‐dock under study. The following characteristics do recur often in academic literature (see appendix A):
- The inbound and outbound trailer fleet is well dimensioned with interchangeable standard trailers having the same capacity
- Handling operations of all trailers take a very similar amount of time - Dock doors are identical (all trucks are able to dock at all dock doors) - Inbound and outbound trailers are operated separately
- All trailers have equal priorities (except for literature with the earliness/tardiness objective) - All empty inbound-‐ and outbound trailers are instantaneously replaced or have fixed
changeover times
- Dock doors are typically dedicated for inbound or outbound activities
- All inbound and outbound trailers are available at time zero, or the arrival sequence of the inbound trucks is known
- Due dates of outbound trucks are typically not taken into account
- Movement times in the dock are negligible (except for the inner transport articles) - Cross-‐docks have unlimited forklift-‐, unload-‐, and load capacity
- The order of loading outbound trucks is not of interest
3. Methodology
This chapter discusses the methodology used for this thesis. The remainder of this chapter is organized as follows: First, the research questions are presented which is followed by a paragraph discussing the scope of the research. The chapter is concluded by an overview of the performed methods.
3.1 Research questions
This thesis is aimed towards exploring the synergy between existing mathematical decision models in literature and LTL cross-‐docking practice and opportunities for improvement. This leads to the following research question:
“To what extend does synergy exist between academic cross-‐docking literature and the operations of the LTL cross-‐dock under study, and what opportunities for improvement exist?”
An extensive literature study is performed to create an overview of the existing cross-‐dock decision models. The research method applied in this study is a single in-‐depth case study approach, which has been used to get a thorough understanding of the operations, characteristics, and constraints of an LTL cross-‐dock carrier in practice. The data retrieved from the literature-‐ and case study is used to assess the similarities and differences between cross-‐docks described in literature and cross-‐docks operated in practice. However, the outcomes of the assessment cannot be directly used to answer the main research question. In order to support answering the main research question the following additional research questions are composed:
– How does the cross-‐dock under study differ from cross-‐docks described in literature and what is the effect on decision modeling?
– Which (aspects of) existing mathematical decision models can be used in the LTL cross-‐dock under study and which changes to physical operations or management approaches are required to make it work?
– What fundamental changes are needed in future research to make the proposed decision models better applicable to the LTL cross-‐dock under study?
3.2 Scope of the research
Academic literature on cross-‐docking can be classified according to the following categories (ordered from strategic to operational) (Boysen and Fliedner, 2010):
– Location of cross-‐docking terminals – Layout of the terminal
– Vehicle routing – Truck scheduling
– Resource scheduling inside the terminal – (Un-‐) packing load into (from) truck
This research is focusing on the operational categories of cross-‐dock literature which can be influenced by the cross-‐dock department. These categories are:
– Assignment of destinations to dock doors – Assignment of destinations to dock doors – Truck scheduling
– Resource scheduling inside the terminal
The other categories are not considered in this research since they are or strategic in nature (e.g. the location of cross-‐docking terminals and the layout of terminals), or they are not applied in the cross-‐ dock under study. The vehicle routing schedule task is performed by the companies planning department, and the (un-‐)packing of shipments (i.e. the activities of repacking inbound pallets to form other shipment compositions on the outbound pallets) is typically not performed at an LTL cross-‐dock.
3.3 Methods
The information for this research is gathered by means of a structured literature review and an in-‐ depth exploratory case study. Insights and conclusions are drawn based on analytical generalization.
3.3.1 Structured literature review
The literature review is performed to search for the main existing decision support models and their characteristics in academic literature. Buijs et al. (2013) are classifying the existing cross-‐docking research according to six distinct cross-‐docking management problem classes. They present their results in a classification table, which is used to select the papers for our research. Because of the scope of our research, only the papers describing operational problem aspects at the cross-‐dock, with the subjects of “dock door assignment”, “truck scheduling” and “resource scheduling inside the terminal”, were selected. The selected papers were reviewed based on the objectives, methods and assumptions in order to create an overview of their decision problem aspects and their characteristics and assumptions. The results of this study are discussed in chapter 2 and presented in a framework which is added in appendix A).
3.3.2 Exploratory case study
The in depth single and typical case study is performed at a relatively large third-‐party logistics provider in The Netherlands. The company has different business units, including an expedition business unit, warehousing business unit and a cross-‐dock terminal. The cross-‐dock terminal is mainly used by the planning department to consolidate shipments to and from international and domestic trucks.
The cross-‐dock has 105 dock doors and handles 175 inbound-‐, and 175 outbound trailers a night on average. In this thesis, the name of the company has been changed to “the case company” due to confidentiality reasons. The data was gathered during a four months internship. The exploratory case study was performed in order to get an insight in LTL cross-‐dock operations in practice, its constraints, and how the case company differs from cross-‐docks described in literature. The results of the exploratory case study are presented in chapter 4. Propositions and recommendations are composed by combining the information from the literature review and the case study. A validation of the recommendations for the case company is performed by a interview with the cross dock manager of the case company.
Empirical data collection
The case study has relied on multiple sources of data namely: observations, informal discussions, and historical data analysis. The observations and informal discussions are collected in field notes. The informal discussions were held on-‐site with employees out of all hierarchical layers of the organization, from material handlers to the cross-‐dock manager, to avoid a one-‐sided picture. Fieldnotes are reported on13 observations and 23 informal discussions (see appendix B).
The historical data analysis is performed using the Key Performance Indicator (KPI) reports made by the cross-‐dock management over the period 1/1/2012 to 31/6/2013. The KPI report includes daily information about:
• the turnover (converted into tons) • the number of shipments
• trailers, shipments per trailer • workforce size (FTE)
• illness
• the percentage of employees versus flex workers • the percentage of trucks meeting their due dates • the turnover per workhour
The main objective of the analysis was to identify the influence of different compositions in production units of the cross-‐dock, e.g., differences between trailers, differences between employees and differences between import and export shipments, at the average productivity of employees in a planning horizon.
3.3.3 Analytical generalization
4. Empirical findings
This chapter presents the results of a single-‐case study in an LTL cross-‐dock. First, the main characteristics of the case company are discussed to provide a better insight into the specific shipment characteristics, the goods flow, the trailer and dock door characteristics and the operational schedule of the case company. Then, the process of the cross-‐dock is explored to give an overview of the cross-‐dock main operations. Third, with the data of this study we attempt to make a generalization of a typical LTL cross-‐dock in practice. Finally, a comparison is made between cross-‐docks described in academic literature and the LTL cross-‐dock under study.
4.1 Characteristics of the LTL case company
This paragraph gives an overview of the case company’s main characteristics. The case company is described in terms of the characteristics of the shipments, the goods flow, the trailer and dock doors, and by the cross-‐dock operation schedule.
4.1.1 Shipment characteristics
Palletized shipments are typically the most common way of transporting loads through a supply chain. Pallets are easy to handle by forklifts and most trailers are arranged on transporting two standard pallets in a row which eases the loading process. However, not all shipments are suited to be transported on pallets and those shipments are typically transported through LTL cross-‐docks. The LTL cross-‐dock tries to bring in also the other more regular shipments from the same supplier, by offering this kind of services. In the case that the cross-‐dock does not offer the service handling of irregular shipments, the perception is that the supplier will look for another logistics service provider for all its shipments. This situation is mainly caused by the very competitive freight market which in general does not allow companies to get only the most profitable and most easy to handle freight; it is more in line with “take what you can get” (Manager cross-‐dock 1, 03-‐06-‐2013and Supervisor cross-‐dock 1, 19-‐6-‐2013).
sizes range from small packages of 10cm x10cm x10cm to large crates of 600cm x 100cm x 90cm and even larger exceptional sized shipments are sent over the cross-‐dock. According to group leader 1 (11-‐6-‐2013) and Manager cross-‐dock 1 (2-‐7-‐2013), the following shipment characteristics are mainly influencing the transfer time of shipments:
- “Heavy shipments”: shipments heavier than 1500kg. These shipments need a heavy forklift for transportation which has to be called in what takes additional time.
- “ADR shipments”: Shipments containing dangerous goods (ADR) have to be staged at a secured ADR location and have to be secured in the trailer against sliding and fall over. Moving ADR shipments causes additional inner transportation time because the forklift generally has to, compared to the regular storage area, make a detour in the terminal to reach the ADR storage area. Additionally, unloading and loading operations take additional time since ADR shipments have to be secured against sliding and fall over.
- “Broad shipments”: Shipments where the shortest size is wider than 1,5 meter. A forklift needs bar extenders to handle this shipment. Before these shipments can be transported the forklift driver has to search for, and pick up bar extenders which takes an additional amount of forklift travel distance and transshipment time.
- “Long shipments”: Shipments where the longest size is longer than 2,4 meter cannot be placed across the trailer since the width of a standard trailer is 2.4 meter. The shipment has to be (un)loaded via the shortest size of the shipment which is in most cases more labor intensive causing additional unloading time. Regularly, a second material handler is needed to unload these kind of shipments
The different handling characteristics of these shipments are visualized in appendix D. Different shipment characteristics also extends the loading time of outbound trucks for another reason. Standard trailers are slightly wider than 2,4 meters. Standardized pallets are 1,2 meters long, so in theory pallets can be loaded two in a row. However, different shipment sizes frustrates the easy loading of pallets since it becomes a puzzle to load all different sized shipments in an outbound truck, especially when the trailer is completely utilized.
4.1.2 Goods flow characteristics