Supply Chain & Transports Optimization

Volkswagen Autoeuropa Case Study

Elizabeth Andreína Ascenção Rodrigues

Department of Engineering and Management, Instituto Superior Técnico

Area: Industrial Engineering and Management

Abstract
The advanced internationalization and evolution of markets and production processes continuously adds
to the complexity of supply chains, along with an increased pressure to improve the sustainability of
processes and optimize their efficiency. In a world that is increasingly demanding to have the right
product at the right time and place, transportation is one of the most fundamental activities within a supply
chain. Given this importance, there is a continuous pursuit of optimization opportunities in this area that
can contribute to an increased efficiency of the supply chain as a whole.
This work addresses the case study of a car manufacturer, Volkswagen (VW) Autoeuropa (AE), with
specific focus on inbound logistics and the challenges the company encompasses when trying to
incorporate transportation costs specificities into transportation and replenishment decisions.
In this paper, two approaches are proposed to solve the problem. First, a manual approach is performed
in order to provide and assess different frequency transportation scenarios, leading to a tool to support
decision-making. Then, a mixed integer linear programming model (MILP) is developed in order to
validate and optimize the current transportation scenarios.

Keywords: supply chain management; transportation decision-making; transportation management;

freight modeling; inbound logistics.

1. INTRODUCTION companies. Without well-developed transportation

Worldwide and across industries, supply chains systems, logistics could not bring its advantages
(SC) are continuously forced to perform into full play. Furthermore, transportation impacts
exceptionally better by the day, customers demand other areas of the company, since usually
higher service levels and competition tightens. transportation costs are diminished by increasing
Within the supply chain, several players need to be the quantity being shipped, so the company needs
taken account. Focused on outcomes, a to build up inventory, which also represents costs.
competitive supply chain capability not only More, and within a just-in-time manufacturing
assures the on-time delivery of component context, companies need to have a very efficient
material, but also the execution of inbound supply and reliable transportation system, since the
to ensure the shipment of quality finished goods at requirements are notified with a short advance,
a competitive price while delivering on-target suppliers need to be ready, otherwise it will be
margins for the manufacturers. necessary to perform special transports, which are,
Therefore, transportation is an indispensable of course, costly.
activity in a supply chain, yet it can account for The problem addressed in this work concerns
more than 50% of total logistics costs (Swenseth & Volkswagen Autoeuropa, which is an automotive
Godfrey, 2002), thus becoming obvious why manufacturing plant of the Volkswagen Group,
optimizing transportation costs is a priority for located in Palmela. The company identified the

1
necessity of optimizing transportation strategies, simultaneously, the overall value (Chopra &
namely concerning the behavior of transportation Meindl, 2007).
rates charged by carriers, since it was identified Supply chains have evolved significantly
that some of the rates structure were not linear and throughout the years, as the term can be traced
had a piecewise structure, with constant values for back to the 1980s (Coyle et al, 2011). From that
different weight values. Being so, there was a point until today, several trends have been gaining
motivation to study if it could be possible to take importance and influenced the way SC are
advantage of these behavior and try to increase managed and designed. On one hand, the
the shipment size and/or decrease the frequency worldwide economic growth has given rise to a
of the shipments, reducing the value per unit and vast consumption of goods, while globalization has
eventually leading to the suppression of some pick- led to a large stream of goods all over the world.
ups, hence saving the cost of some trips to that The production, transportation, storage and
supplier. consumption of all these goods, however, have
The objective of the present paper is to study, in created large environmental problems. In the
terms of costs, the current transportation sense of integrating adequate environmental
scenarios, and their implications in inventory with management into supply chains and its activities,
the goal of proposing alternative scenarios that the concepts of green logistics and green supply
lead to an improvement of the overall efficiency. chain management (GSCM) have emerged, with
To do this two approaches were undertaken: first in the aim to reduce the environmental externalities of
a more manual manner using MS Office Excel and logistics operations. (Green Logistics, 2016). On
the second, with a Mixed Integer Linear the other hand, there is a shift in the traditional
Programming (MILP) model proposition. view of SC towards a network approach meaning
The paper is structured as follows: in Section 2, that companies must find a way to improve
relevant literature on the design and planning of communication and information flow, thereby
supply chains, transportation decision making and converting the traditional supply chain into a
transportation network design is presented. In flexible and real-time supply network. (Ladd, 2004).
Section 3, a complete problem statement is Transportation in Supply Chain
presented. The manual approach and Transportation Role and Decisions in Practice
mathematical model are presented in Section 4. In
Section 5 the approaches are applied to different Transportation is one of the most structuring
scenarios and the results are compared. A activities of Logistics and it’s responsible for a big
sensitivity analysis is conducted in order to study share of the logistics costs, around one to two
the influence of the variation of certain parameters. thirds (Ballou, 2006). The goal of making the SC
Finally, in section 6 some conclusions are drawn. more agile also transfers a big share of the costs to
the transportation sector, while trying to decrease
2. LITERATURE REVIEW costs in inventories and warehouses. With
Supply Chain and Supply Chain Management transportation being so essential in a SC, it cannot
be managed in an isolated manner, since options
Today, individual firms no longer compete as about the mode and the type of transportation
independent entities with unique brand names, but solutions have a significant impact in the cost
rather as integral parts of a network. As such, the structure and in the company’s ability to react to
ultimate success of a firm will depend on its demand, as well as noteworthy repercussions
managerial ability to integrate and coordinate the throughout the whole SC. Only a good
intricate network of business relationships coordination between each component would bring
necessary to deliver products to customers the benefits to a maximum (Tseng, 2005).
(Lambert & Cooper, 2000). When it comes to transportation decisions there
Different authors present diverse definitions, found are some key trade-offs to consider:
in the literature, that intend to explain the Supply  Transportation and inventories - Transportation
Chain concept. In shorter words, Ballou (2004), time of goods has an increased impact on
states that the term supply chain refers to all of inventory, whether it is inventory in transit, cycle
those activities associated with the transformation inventory or safety stock. Thus, transportation
and flow of goods and services, including their decisions should always incorporate the influence
attendant information flows, from the sources of it has on stock costs.
raw materials to end users, with the traditional  Time versus space - Also, frequency is one of the
objective of minimizing the total supply chain cost most relevant and impactful characteristics, with
to meet fixed and given demand. However, this great importance in costs. As a matter of fact, a
should be performed whilst maximizing,

2
 high supply frequency shows a great response Modelling and defining transportation costs
capacity from the supplier, but can translate into Transportation costs can greatly differ dependent
higher transportation costs. Space availability in of a series of variables. Authors in scientific
the destination facilities is also one factor that research have tried to develop different
needs to be considered when deciding approaches to tackle some of the possible specific
transportation modes and consequent shipment differences and price structures.
sizes (Crespo de Carvalho, 2010). Guo, Wang & Zhou (2015) consider the
In sum, a trade-off between responsiveness and transportation problem with uncertain costs and
costs (transportation, inventory, space, among random supplies. Sahin, et al. (2009) develop a
others) must always be evaluated. model for analyzing total transportation costs, for
When designing a transportation network, initial cargo or passengers, by breaking down the total
transportation decisions are seen as strategic, with cost into: capital, fuel, lubricants, operational and
a broader focus on the overall supply chain maintenance costs. More specifically into
transport system and then the decision-making transportation and the integration of piecewise
scope becomes increasingly operational in nature, linear cost functions into network design decisions,
focusing on activities that implement the overall Sheng, Dechen & Xiaofei, (2006) investigate the
system decisions (Stank & Goldsby, 2000). In transportation problem with discontinuous
general, transportation modelling and network piecewise linear cost function (TPDPLC). They
design and decisions face some commonly argue that TPDPLC can be easily formulated and
encountered problems, as mode selection, carrier solved as a MILP, and theoretically any general
selection, carrier routing, fleet sizing, and vehicle MILP could be used to solve this kind of problem,
scheduling and shipment consolidation. however due to computational requirements,
Integrated transportation decision making metaheuristics methods can lead to good
The importance of transportation management with solutions, and the genetic algorithm proved itself to
a broad concern for the entire SC has been be the most suitable for their work. Görtz & Klose
repeatedly stressed, as much as it has been (2007) study the single-sink fixed-charge
highlighted by the scientific community. In Crainic transportation problem (SSFCTP), which consists
& Laporte (1997), authors aim to identify the main in finding a minimum cost flow from a number of
issues in freight transportation planning and supplier nodes to a single demand node. In this
operations, with a classic three decision levels transportation problem, shipping costs comprise
approach (strategic, tactical and operational). costs proportional to the cargo as well as a fixed-
Stank & Goldsby (2000), develop a framework for charge. Authors refer that, though it is a very
transportation decision making in an integrated SC, important relaxation from the fixed-charge
since they identified, at the time, the lack of transportation problem, it has not attracted much
conceptual research regarding the role of attention in the literature, with only a few authors
corporate transportation function in an integrated actually developing meaningful insights. The
SC. Regarding network flows, Kim & Pardalos applications for the SSFCTP were first referred in
(1998) consider the fixed charge network flow 1996, according to the authors, and include the
problem (FCNFP), and propose a new approach areas of manufacturing and transportation, namely
for solving the general capacitated (or in supplier selection, product distribution/fleet
uncapacitated) FCNFP by adapting an economic selection and process selection. Christensen
viewpoint of the fixed cost. Authors acknowledge (2013) composes his thesis by addressing three
that this problem has many practical applications different problems with commonalities between
including transportation, network design, them, since all of them are related to Network
communication and production scheduling, and Design and consider a piecewise linear cost
although many exact methods have been structure. The beginning of the author’s work
developed to solve the FCNFP, their computational reflects an overview of different facility location
requirements increase exponentially with the size models, using non-standard production and/or
of the problem, so heuristic approaches are transportation cost function such as the piecewise
needed to solve sub optimally large-scale linear cost function, with the aim of categorizing
problems, and most solution approaches until that existing literature, which led to the conclusion, as
time included branch and bound techniques and previous authors, that the area of non-standard
vertex ranking procedures. transportation costs had only been sparsely
treated in literature. Furthermore, he focuses on
the single-sink fixed-charge multiple choice
transportation problem (SSFCMCTP) by

3
presenting an efficient dynamic programming distribution side from producers to customers in
algorithm to solve it. Then, the work leans on more terms of sustainable supply chain management
general capacitated facility location with piecewise with a green multi-objective optimization model that
linear transportation costs problems. minimizes CO2 emissions from transportation and
Modelling transportation and inventory costs total costs in the distribution chain. Existing
Dhakry & Bangar (2013), stress the fact that there literature on economic and environmental concerns
are two important topics in the supply chain that along with social repercussions is rare, hence is
contribute to the total cost of the supply chain important to remark the study by Ramos, Gomes &
network: transportation and inventory costs. With Barbosa-Póvoa (2014) in which they aim to provide
that in mind, companies can achieve significant support at tactical and operational decision making
savings by considering these two costs at the levels of reverse logistic systems planning, while
same time rather than trying to minimize each considering economic, environmental and social
separately. Three in-stock strategies are studied in objectives. Authors then model the problem as a
their work – flow through, regional and single DC multi-objective, multi-depot periodic vehicle routing
central stock – and they develop a simple problem with inter-depot routes.
transportation-inventory model to compare the total 3. PROBLEM STATEMENT
costs of each strategy. A distribution model is also
formulated as a non-linear integer optimization This challenge began with the identification, from
problem, and due to the non-linearity of the Autoeuropa, of a particular behavior regarding
inventory cost in the objective function, two transportation inbound rates for certain suppliers in
heuristics and an exact algorithm are proposed. certain countries. This behavior, as explained
Also, Swenseth & Godfrey (2002) work aims at before, presented characteristics of a non-
incorporating transportation costs into inventory decreasing piecewise function, with constant
replenishment decisions, as the need to develop segments and others that increased in a fairly
models with appropriate representation of linear manner according to the cargo weight.
transportation considerations is further enhanced. Having identified the existence of these unusual
So they propose a solution procedure, based on structures, the objective was to narrow the
the economic order quantity model (EOQ) and countries in which that kind of cost structure was
further modifications, that systematically solves present and consequently consider the suppliers
inventory replenishment problems, using a present in those countries. Furthermore, it was
classification of shipping decisions according to intended to analyze how those rate profiles could
whether shipments can (or are) declared as full be used in an advantageous way for the firm,
truck loads or not. Zeng, Hu & Huang (2013) especially when cargo’s taxable weight fell into the
analyze the transport allocation problem in a constant price segments, since the comprehensive
multimodal transportation environment for knowledge of those prices could allow to increase
automotive logistics, by using a genetic algorithm cargo sizes until the limit of the segment, and thus
that can minimize logistics costs. exploring scenarios of less frequent shipments. Of
course transportation scenarios have impacts on
Incorporating sustainability into transportation
inventory, since the reduction of frequencies
decisions
implies an increase in order quantities and
Due to the current growing concern with therefore stock build-up.
environmental issues and sustainability in the SC,
there is also an increase in literature research 4. APPROACHES EXPLANATION
focus to attempt to integrate sustainability and In order to solve and propose improvements on the
green logistics aspects into transportation current scenario, two approaches were developed:
decisions. Dekker et al. (2012) present a review a manual approach exploring excel and an
that highlights the contribution of Operations optimized approach.
Research (OR) to green logistics. Ji, Wu & Zhu
4.1. Manual Approach
(2015) address the issue of eco-design for
transportation in SSCM, using and extension of First, a manual analysis was performed for each
Data Envelopment Analysis (DEA) and its supplier, in which different scenarios were studied
combination with a developed tractable algorithm. by changing the frequency of orders to a biweekly
Tang et al. (2015) evaluate the issue of cutting or a monthly scenario. Also, given the high level of
emissions by reducing shipment frequency within a detail associated with this approach it was also
periodic inventory review system environment. analyzed for some suppliers a scenario of quantity
Validi, Bhattacharya & Byrne (2014) look at the changes, with the goal of increasing some units in

4
every order in order to accumulate stock – and Mathematical formulation
simultaneously take advantage of the constant Sets/Indices
step of the function in which the current order was
Set of suppliers i
– and reach a week in which it was not necessary
Set of time periods t
to place an order.
Set of products p
This approach was performed using the
potentialities of Microsoft Office Excel, and it had Set of possible lots l
under consideration the forecasted deliveries, Set of products p supplied by supplier i
packaging type, packaging dimensions, number of Set of possible breakpoints j in the
parts per packaging and product weight. The piecewise transportation cost function.
forecasted orders allowed to understand what the Α Set of countries α
necessary quantities were and the remaining data Parameters
led to the total taxable weight. For the other - Price per product p
proposed scenarios – biweekly, monthly, and - Opportunity cost per product p
quantity increases -, the forecasted deliveries were
- Opportunity rate
altered to reflect those delivery scenarios and
calculations to obtain total taxable weight were - Initial stock per product p
performed using those new values. Furthermore, - Safety stock per product p
impact on inventory was assessed by analyzing - Maximum allowable capacity
the material requirement planning for each product, - Number of unit products per product
as to know the quantities of product that were package
required to feed the line per week, and by - Taxable weight factor per product p
performing an inventory balance calculation, to - Cost parameters (y-axis) for piecewise
determine the increase in the average stock per cost function per supplier i, in breakpoint
week that each scenario would provoke. The j and country α
manual approach was then automatized into an
- Weight parameter (x-axis) for piecewise
Excel tool to aid Autoeuropa transportation
planners. cost function per supplier i, in breakpoint j
and country α
4.2. Optimization Approach
Variables
Regarding the optimization approach a
mathematical model was developed from scratch. - Quantity of product p to be ordered in time t
In this, different products can be delivered by - Transportation cost per supplier i in time t
different suppliers into a single destination point. - Cargo weight value per supplier i in time t
Products – raw materials and parts – can only be - Lot sizes l possible to be ordered per
supplied by one supplier, yet one supplier can be product p
responsible for the delivery of more than one - Number of packages per product p in
product. time t
Generically, the problem can be described in the - Taxable weight per product p in time t
following summarized form:
Given: - Taxable weight per supplier i in time t
- Transportation cost function - Stock quantity per product p in time
- Material Requirement Planning for each period t
time in each time period - Total stock cost per product p in time
- Established maximum capacity of vehicles period t
- Products information: prices, etc. - Total stock cost per supplier i in time period
- Stock opportunity cost t
- Positive variable, coefficients for the
Determine:
piecewise cost function linearization, for
- Optimum quantities from supplier to
supplier i for country α, breakpoint j, in
destination
time t.
- Optimum delivery frequencies
Binary
With the goal of minimizing transportation and - Binary variable, 1 if lot size l per product p
inventory costs. in time t is selected; 0 otherwise

5
 - Binary variable, 1 if the breakpoint j of the selected in time t; 0 otherwise.
cost function for supplier i in country α is

Objective Function
(1)
Subject to:
(2)

(3)
(4)

(5)

(6)

(7)

(8)

(9)

(10)

, (11)
+ ,
(12)

(13)

(14)
(15)

(16)
(17)
(18)

(19)

(20)

The model presented can be explained as follows: function of the supplier and the cost structure that
First, the objective function (1) minimizes the total has been discussed previously; then the second
costs of the transportation activity, with its first term term regards inventory costs, so it is the total cost
relating to the transportation costs directly, as a

6
for the analyzed period of time of the stock build- week minus the requirements of that material.
up. Equation (14) guarantees that stock levels are
Equation (2) relates to the quantity that should be never below the defined safety stock for the
ordered, of each product in each moment, stating product.
that the quantity to order can only be one of the Equation (15) is related capacity limitations.
options that belongs to the set of lots, which define However, as explained before, this is not an actual
the fixed quantity sizes. Equation (3) uses the physical constraint from the truck, but it regards the
quantity to be ordered divided by the parameter of defined scope of this project in terms of maximum
the number of parts per package in order to taxable weight that is studied to avail from the
translate the quantity into package units. stepwise cost structures.
Equations (4) and (5) are calculations for the Equation (16) models the calculation for the
taxable weight. The first one is per product and opportunity cost in which inventory is valued as the
shows how the taxable weight depends on the result between the opportunity rate chosen by the
number of packages and the factor – which is the company and the price of the product.
highest unitary value between weight and volume Equations (17) and (18) are support calculations
as explained in previous chapters. that show how the inventory cost is calculated per
The second formula shows the summation for product and then per supplier, by summing on the
suppliers, according to the set of products of the set of products belonging to the supplier. Finally,
supplier. For Portugal, these values do not equation (19) states that the total taxable weight
interfere with transportation costs, however they per supplier is the value for the abscissas in the
are fundamental for the model for suppliers in cost function. Equation (20) equals the abscissas
Spain and France. to the taxable weight.
Equation (6) relates to equation (1), as it is a The final model results in a mixed integer linear
constraint that forbids that more than one lot size is programming model (MILP) that was implemented
chosen for the quantity to be ordered for each in GAMS.
product in each time period.
5. RESULTS
Equation (7) and equation (8) represent the
ordinates and abscissas values for the The approaches explained previously were applied
transportation cost calculation, as it has to be to the case study. First, the manual approach with
modeled following the linearization of piecewise three frequency scenarios was considered:
function explained in the previous section biweekly, monthly and quantity changes.
Equation (7) stands for the transportation cost, Following, the mathematical model was applied to
which is the sum of the possible ordinate optimize and analyse the current frequency
breakpoints values multiplied by the lambda delivery situation at AE.
parameter that defines the two breakpoints “active” 5.1. Manual Results for Portugal
in that time period. The same logic follows for
The impact – when compared with the original
equation (8), however represents the x-axis scenario - of biweekly and monthly scenarios
values of the breakpoints, meaning total taxable analyzed for portuguese suppliers are shown in
weight per supplier. table 1 and 2. It can be seen that these scenarios
Equation (9) is the statement that guarantees that lead to transportation savings, but that the final
the sum of parameters “lambda” equals one for result for most of them is negative due to the
each supplier in each time period. Equation (10) excessive cost of inventory these scenarios
guarantees that only two binary variables are encompass. Some suppliers were not analyzed on
active and consequently only two breakpoints are biweekly scenarios due to the fact that their current
selected in the cost function to define costs in that frequency scenarios were already very similar to
period of time. Thus, equation (11) guarantees that biweekly.
those two breakpoints are consecutive by
Table 1 - Manual Results for Biweekly Scenario in Portugal
preventing non-consecutive points of being active
at the same time.
Equations (12) to (13) concern inventory
constraints. Both equation (12) and (13) are
material balances, for the first time period and the
following ones, respectively. Here is stated that
inventory for a product during a week is the sum of
stock from the previous week (or initial stock for
equation 12) plus the quantity that is received that

7
Table 2 - Manual Results for Monthly Scenario in Portugal product was increased and the remaining ones still
needed to be ordered at the same previous weeks.
5.3. Manual Results for France
The results of both scenarios analyzed for this
country when compared to the original are
summarized in Table 6 and 7.
5.2. Manual Results for Spain Table 6 - Manual Results for Biweekly Scenario in France
Table 3, 4 and 5 show the impact of the three
scenarios assessed for Spain. Also, some
suppliers were not analyzed on monthly scenarios,
as the junction of cargo would largely exceed the
constraints defined for the scope of this work.
Therefore, a third scenario arose – the quantity
changes scenario – with the goal of increasing
some units in each order as to avail from the
constant step of the piecewise cost function and
then reach a week were it would not be necessary Table 7 - Manual Results for Monthly Scenario in France
to order, as needs would be covered by built stock.
Table 3 -Manual Results for Biweekly Scenario in Spain
Impact Biweekly Monthly Quantity changes
(WK 8-20) Transportation Inventory Rental Total Transportation Inventory Rental Total Transportation Inventory Rental
Supplier 1 - 127,58 € 1 863,25 € 2,84 € 1 738,52 € - 97,39 € 32 876,73 € 4,5
Supplier 2 - 423,63 € 1 554,44 € 8,33 € 1 139,14 €
Supplier 3 - 39,49 € 4 615,81 € 2,89 € 4 579,21 € 115,23 € 224,46 € 0,0
Supplier 4 - 156,38 € 294,69 € 1,01 € 139,32 € - 158,63 € 7 395,87 € 1,1
Supplier 5 9,51 € 5 912,65 € 0,5

Table 4 - Manual Results for Monthly Scenario in Spain

Results show, as with the previous countries, that
these scenarios allow for transportation savings for
almost all suppliers. However, only one of the
suppliers has an overall positive situation, which is
supplier 2.
Furthermore, it can be seen that supplier 4 has
increased transportation costs in the monthly
scenario, explained by the fact that cost functions
Table 5 - Manual Results for Quantity Changes Scenario in
Spain
in this country have small constant steps and thus
the increase of quantity in this scenario leads to
the cost values jumping onto the following piece of
the piecewise function, increasing total
transportation costs, though unit costs are
reduced.
5.4. Optimization Approach Results
Overall, the mathematical model aimed to provide
The observation of the tables shows that there are the optimum quantities to be ordered from a
positive savings in transportation for mostly all supplier, having under consideration the possibility
suppliers, but similarly to the portuguese case, of availing piecewise transportation cost structures,
inventory precludes the potential overall savings. It leading to reduced frequencies that could cost the
is important to notice that for some suppliers, the same per frequency and thus reducing
quantity changes scenario did not prove to be transportation costs. At the same time, the model
advantageous. This scenario was proven to be considered the minimization of inventory costs,
benefitial for suppliers with only one product, which are highly dependent on the quantities to be
because for the others there were either a jump in ordered. Therefore, there was a challenge in trying
the cost function – increasing costs – or only one to obtain the best combination of orders from each
supplier and for each product while having the

8
 Table 10 - Scenarios Comparison for Portugal
capacity to assess the trade-off between two
fundamental costs for logistics. Given so, overall TOTAL COSTS PORTUGAL
results for the countries under analysis are ORIGINAL 8.485,52€
summarized in Table 8 and 9. BIWEEKLY 8.407,33€ -1%
MANUAL
Table 8 – Total Transportation Results (Model) MONTHLY 12.932,99€ 52%
Transportation PORTUGAL SPAIN FRANCE MODEL GAMS 5.371,78€ -37%
Original Quantity 123 488 176 960 326 770 Table 11 - Scenarios Comparison for Spain
GAMS Quantity 112 976 -8,5% 176 210 -0,4% 303 788 -7,0% TOTAL COSTS SPAIN
Original Cost 759,53 € 3 084,89 € 7 445,24 € ORIGINAL 20.338,61€
GAMS Cost 569,65 € -25,0% 3 040,27 € -1,4% 6 603,90 € -11,3% BIWEEKLY 20.807,61€ 2%
Table 9 – Total Inventory Results (Model) MONTHLY 23.630,03€ 16%
MANUAL
Inventory PORTUGAL SPAIN FRANCE QUANTITY
24.370,10€ 20%
Original Quantity 590 519 327 264 670 271
CHANGES
GAMS Quantity 302 379 -48,79% 310 738 -5,05% 505 007 -24,66% MODEL GAMS 15.071,68€ -26%
Table 12 - Scenarios Comparison for France
Original Cost 85 844,34€ 191 707,96€ 346 067,59€
Cost GAMS 53 357,00€ -37,84% 133 682,35€ -30,27% 284 760,14€ -17,72% TOTAL COSTS FRANCE
From both tables is possible to conclude that the ORIGINAL 38.591,32€
mathematical model allows to significantly improve BIWEEKLY 48.725,71€ 26%
MANUAL
quantities and costs for all countries. MONTHLY 81.184,74€ 110%
This is however an overall result, and there are in MODEL GAMS 32.232,31€ -16%
fact some suppliers within the countries that On the other hand, the mathematical formulation
increased costs in one of the areas – whether proved to significantly optimize the current
transportation or inventory – due to the fact that scenario by values between 16% and 37%, as
the model tries to balance the sum of both costs. seen in the last row of the above tables, thus
Given the fact that inventory valorization is fully revealing that the current methodology of orders
proportional to the cost of the product, the rationale forecasting might not be the most advantageous
supporting these results is sustained on the high for the company.
costs some of the products have and,
5.6. Sensitivity Analysis
simultaneously, on the very reduced costs some
other possess, which allows the model to increase During the development of this work, there was a
the quantity of those that are not that expensive – constant preoccupation regarding the parameter of
possibly reducing transportation frequencies and opportunity cost considered by the company. This
costs - and therefore will not impact as strongly value is directly related to inventory valorization,
the final result as well as decrease the quantity – since the calculation is based on the multiplication
possibly also augmenting frequencies and of the average stock level, the product price and
resulting in higher transportation costs - , and the opportunity cost. The value used by AE, is 9%,
consequently the stock costs, of those that are meaning that the company assumes that the
more costly. money that is tied up in inventory could result in an
interest rate of 9% if invested in an alternative
5.5. Scenarios Comparison
financial application. Given the current market and
Overall, the two different approaches applied to the financial conditions, this does not seem like a
case study led to different results that are feasible scenario, which is why this parameter is
summarized in Tables 10, 11 and 12 for each further studied in this section, in order to assess
country. This tables show the total costs for all how its reduction, to more real values, could
scenarios analyzed under the current conditions impact the analysis previously presented.
(opportunity cost) and its percentage comparison The parameter was submitted to 1% reductions,
to the original scenario – green and negative from 9% until 3% Results show this parameter
means savings and red and positive means precludes all the manual scenario analysis and it
increased costs. From its observation, along with also reduces the potential of savings from the
the results that were shown in the previous mathematical model. Given the fact that it is out of
sections, it is possible to conclude that, though sync with the current market values, it is advised
manual scenarios have very positive transportation that the company reconsiders its inventory
savings for mostly all suppliers, their total results valorization policy, as it might be hampering great
are deeply jeopardized by the increased inventory savings.
costs, resulting in final values that are not so
advantageous for all suppliers.

9
6. CONCLUSIONS with a tool that can aid transportation managers in
taking more efficient decisions and incorporating all
In this work the inbound logistics of an automotive the knowledge that was now understood and
manufacturing plant regarding the replenishment identified by the development of this work in the
decisions with incorporation of transportation and company. Furthermore, an academic paper will be
inventory costs is analysed. written and published in the following year in an
Two approaches were developed: a manual approach academic magazine
that led to the development of an Excel tool; and an
optimization model which integrates simultaneously 7. REFERENCES
transportation and inventory costs in order to provide Ballou, R. H. (2004). Logística: Administración de la Cadena de
the best schedule of orders that minimizes the impact Suministro (5th ed.). México: Pearson Educación.
Ballou, R. H. (2006). The evolution and future of logistics and supply
on both variables. chain management. European Business Review, 19(4), 332 – 348.
Results obtained show that there is potential for Coyle, J., Novack, R., Gibson, B. & Bardi, E. (2011) Transportation: A
transportation savings and that the current supply chain perspective. 7th edition, South-Western College Pub.
Crainic, T. G., & Laporte, G. (1997). Planning models for freight
forecasting method at the company might not be the transportation. European Journal of Operational Research, 97(3), 409–
most advantageous. However, the company needs to 438.
confront the improved results – which are obtained Crespo de Carvalho, J. (Coordinator) (2010). Logística e Gestão na
Cadeia de Abastecimento. Portugal : Edições Silabo.
having a series of assumptions and simplification Chopra, S., & Meindl, P. (2007). Supply Chain Management: Strategy,
under consideration – with real constraints. planning and [Link] Saddle River, NJ: Pearson Prentice
As future work development a few suggestions arise Hall
Dekker, R., Bloemhof, J., & Mallidis, I. (2012). Operations Research for
regarding: broadening the sample of suppliers, green logistics - An overview of aspects, issues, contributions and
relaxing the capacity constraints, verify the challenges. European Journal of Operational Research, 219(3), 671–
679.
applicability of scenarios with suppliers, modelling Dhakry, N. S., & Bangar, P. A. (2013). Minimization of Inventory &
environmental impacts as an objective variable. To Transportation Cost of an Industry - A Supply Chain Optimization.
conclude, though there are aspects that can be International Journal of Engineering Research and Applications, 3(5),
96–101.
improved in future steps, it is expected that this work Green Logistics - Research into the sustainability of logistics systems
stands as a meaningful contribution to the academic and supply chains. (n.d.). Retrieved May 30, 2016, from
community about the study of piecewise [Link]
Guo, H., Wang, X., & Zhou, S. (2015). A Transportation Problem with
transportation cost functions in a real example and Uncertain Costs and Random Supplies. International Journal of E-
that the developed work can be of extensive Navigation and Maritime Economy, 2, 1–11.
usefulness for the decision making process of Ji, X., Wu, J., & Zhu, Q. (2015). Eco-design of transportation in
sustainable supply chain management : A DEA-like method.
Autoeuropa’s transportation planners. Transportation Research Part D: Transport and Environment, 48, 451-
459.
Original contributions and potential impact Ladd, D. (2004) From Supply Chain to Supply Network: A New Paradigm
in Customer Service and Productivity. Information management.
It is remarkably important to notice that this work was Retrieved May 30, 2016, from
developed completely by the author, with all the tools [Link]
specialreports/20041026/[Link]
and developed model being built by the author from Lambert, D. M., & Cooper, M. C. (2000). Issues in Supply Chain
scratch, having as inspiration the existing literature Management, 83, 65–83.
but with the concern of developing a model that could Ramos, T. R. P., Gomes, M. I., & Barbosa-Póvoa, A. P. (2014). Planning
a sustainable reverse logistics system: Balancing costs with
be general enough to constitute an important environmental and social concerns. Omega (United Kingdom), 48, 60–
contribution for the academic community as well as 74.
sufficiently incorporative of the case study Sahin, B., Yilmaz, H., Ust, Y., Guneri, A. F., & Gulsun, B. (2009). An
approach for analysing transportation costs and a case study.
specificities. This work intends to fill the void that was European Journal of Operational Research, 193(1), 1–11.
perceived in the literature research regarding the Sheng, S., Dechen, Z., & Xiaofei, X. (2006). Genetic Algorithm for the
combination of the single sink transportation problem Transportation Problem with Discontinuous Piecewise Linear Cost
Function. International Journal of Computer Science and Network
and the transportation problem with discontinuous Security, 6(7), 182–190.
piecewise linear costs, while also incorporating the Stank, T. P., & Goldsby, T. J. (2000). A Framework for Transportation
Decision-Making In An Integrated Supply Chain. Supply Chain
essential concerns with the impact transportation Management: An International Journal, 5(2), 71–78.
decisions have in the remaining logistic activities, Swenseth, S. R., & Godfrey, M. R. (2002). Incorporating transportation
more specifically inventory repercussions. Therefore, costs into inventory replenishment decisions. International Journal of
Production Economics, 77(2), 113–130.
the developed work provides a solid basis for further Tang, S., Wang, W., Yan, H., & Hao, G. (2015). Low carbon logistics :
studies in the area of replenishment decisions that Reducing shipment frequency to cut carbon emissions. Intern. Journal
can minimize costs while simultaneously considering of Production Economics, 164, 339–350.
Tseng, Y. (2005). The role of transportation in logistics chain. Eastern
stock costs and transportation costs, with the very Asia Society for Transportation Studies, 5, 1657–1672.
important characteristics of the transportation cost Validi, S., Bhattacharya, A., & Byrne, P. J. (2014). A solution method for
structure not being linear and completely proportional a two-layer sustainable supply chain distribution model. Computers &
Operations Research, 54, 204-217.
to cargo weight, as most of the literature case studies Zeng, T., Hu, D., & Huang, G. (2013). The Transportation Mode
consider. Also, and with high importance, is the fact Distribution of Multimodal Transportation in Automotive Logistics.
that the company under study, VW AE, was provided Procedia - Social and Behavioral Sciences, 96, 405–417.

10