A system dynamics model for strategic management of spare
parts in closed-loop supply chains
Marcus Schroter* and Thomas Spengler**
* phone: +49 531 391 2204; marcus.schroeter@ tu-bs.de
** phone: +49 531 391 2201; t.spengler@ tu-bs.de
* ** Department of Production Management, Technical University of Braunschweig,
D-38106 Braunschweig, Germany
ABSTRACT
The strategy to recover components from discarded electrical and electronic equipment to
obtain spare parts is promising, especially during the final service phase. In that phase, the
original product is no longer produced and the sources of new parts are often limited.
Controlling those closed-loop supply chains is challenging. Decision makers have to choose
when to acquire discarded equipment, when to recover used parts, and when to produce new
parts. We developed a generic system dynamics model that provides a test for various
proposed policies to control closed-loop supply chains with parts recovery and spare- parts
supply.
Keywords: Closed-Loop Supply Chain Management, System Dynamics, Spare parts, Parts
recovery
1, INTRODUCTION
The concept of extended product responsibility is getting increasingly important in
electronics industry. New laws and regulations oblige original equipment manufacturers
(OEM) of electrical and electronic equipment (EEE) to take back and recover their products
at the end of their useful lives. A far-reaching attempt at regulation is the European Unions
directive on waste electrical and electronic equipment (WEEE). Similar regulations can be
found in Japan, Korea, and Switzerland. Therefore, the processes of acquisition/collection,
testing/grading, reprocessing and redistribution expand the traditional value chain of
producers of EEE and thus closed-loop supply chains (CLSC) appear (Fleischmann 2003).
Instruments to plan and control these emerging processes will have to be developed and
provided with relevant information. To take advantage of potentials that will arise from
closing loops in supply chains, OEMs may try to use product retums as an additional source
to replenish materials, components, or products. We concentrate in this paper on the
applicability of component recovery strategies to obtain spare parts that are in their final
service phase for OEMs of durable EEE. The final service phase begins when the original
product is no longer produced. Since producers of EEE must assure the customer of spare-
parts supply at least for the average lifetime of the equipment, parts recovery gets promising
because the procurement and the manufacturing of spare parts often becomes difficult when
they are in their final service phase (Ihde, Merkel, and Hanning 1999) (Fleischmann, van
Nunen, and Graeve 2002) (Teunter, and Fortuin 1999) (Hesselbach, Mansour, and Graf
2002).
Thus, at the beginning of the final service phase OEMs frequently predict the entire
demand during the remaining service time and manufacture a final lot of that size.
Determining final orders accurately is complicated, because a long-term forecast over several
years is needed and a number of intangible factors must be considered (Fortuin 1980) (Schulz
1977). In practice, the authors found deviations of about 30-100 per cent over a period of
only 30 per cent of the total remaining service period. Especially, the underestimation of
spare- parts demand is extremely undesirable, because producers of durable EEE must find
altemative sources to maintain spare-parts supply. To meet the demand, firms often have to
redesign the specific part. This procedure can be costly, can take several months, and is
therefore extremely undesirable for OEMs. Furthermore, the running out of spare parts occurs
usually after a considerable time delay after producing the final stock. Because of personnel
fluctuation in spare- parts- divisions during that period, the responsibility for spare parts in the
final phase is often unclear. Therefore, monitoring of final stocks is often neglected.
It is proposed that parts recovery from discarded equipment is a promising strategy to
mitigate the risks of final order strategies, since it raises the operational flexibility of final
order strategies (Schroter, and Spengler 2004) (Spengler, and Schroter 2003) (ZVEI 2002)
(Teunter, and Fortuin 1999) (Hesselbach, Mansour, and Graf 2002) (Hagen, Iding, and
Sudhoff 2000). Hence, adequate closed-loop supply chains have to be designed. The type of a
CLSC we considered in this paper is depicted in figure 1.1.
Collection | Selection bee Reprocessng
L. Disposal
-{ Spare Parts flows in the original chain |
== >| Equipment flows in the recovery chain |
= «> Parts flows in the recovery chain
Disposal
ayer
Figure 1.1: Closed-loop supply chain (similar to Fleischmann (Fleischmann 2001))
In such a CLSC reverse logistics and spare parts logistics processes have to be coordinated,
which is challenging since the problem is dynamically complex. The aim of the paper is to
design robust policies for an integrated management of component recycling and spare- parts
supply. To achieve this objective, we developed a strategic- management tool grounded on the
theory of system dynamics. The remainder of the paper is organized as follows. In section
two, we propose a generic system dynamics model of a closed-loop supply chain for capital
goods - for example, medical equipment, copiers, or prepress and printing systems. The
model will be used in section three to explain the difficulties that decision makers have in
managing parts recovery and spare-parts supply in the final phase, and to test various
strategies OEMs of EEE and recovery companies might pursue to improve the functioning of
the CLSC. Based on this, conclusions are derived in section four.
2. THE MODEL
In this section, we present a descriptive overview of a generic system dynamics model to
analyze and assess various policies to control the CLSC. The model allows us to understand
the dynamic behavior of the CLSC from investigating its physical and information flows, its
organizational structure, and its decision mules, used within the CLSC.
The model consists of four sectors. Figure 2.1 depicts a model structure overview
(Spengler, and Schroter 2003).
Spare-parts supply
Spare-parts demand
Production and
distribution
Aging process of
equipment
Il
I Revenues
and costs
Returned
Reverse logistics
and component
recovery
Economic assessment
Figure 2.1: Model structure overview
2.1 Aging sector
The number of equipment sold is an exogenous input to the aging sector of the model
(equipment sales). Within the aging sector, the equipment in use is modeled as an aging
chain, consisting of cohorts for each year of age of the equipment. We refer to Sterman
(Sterman 2000) for the details of modeling aging chains and to Zamudio- Ramirez (Zamudio-
Ramirez 1996) who uses a similar approach to model an aging chain of automobiles. To
consider product retums we modeled obsolescence rates that are the fraction of equipment
that will no longer be used, because of fashion, functional or economic obsolescence. We
assume that the firm cannot control the obsolescence of equipment. Thus, the obsolescence
parameters are exogenous inputs to the model. To consider the fact that users often store
obsolete devices instead of directly passing them on to recycling companies, we modeled a
“virtual” stock, where the units of obsolete equipment are flowing. Once they are given back
to the CLSC, reusable parts can be extracted. Furthermore, users often make use of alternate
disposal options, for example passing them to independent recovery companies. This user
behavior causes additional difficulty in estimating product retums.
Since some users may want to keep on using their non functional equipment after
repairing it, we defined age-dependent repair parameters. These parameters determine spare-
parts demand.
2.2 Spare-parts supply sector
The spare-parts supply sector models how the firm fulfills spare parts orders, and how it
produces and stores spare parts in the final phase. We use the policy structure diagram shown
in figure 2.2 to provide an overview of the spare-parts supply, and the reverse logistics and
recovery sector of the model. Rounded rectangles represent policies, decision rules, or
organizational subunits that are not shown in detail (Sterman 2000).
[Frage] inl Order
Production
Capacity
Production
Scheduling
Product
Acquistion —
Policy Scheduling
ing Secor KoTORSOE Tockofreumed| _y_
Toa Equipment] | Faaien Eauipmend as
Obsolescence [are-Bacl Rate]
ea a
Figure 2.2: Simplified structure of the spare-parts supply sector, and the reverse logistics and recovery sector
A part of the spare-parts supply sector takes pattern from Sterman (Sterman 2000). We
expand it by the determination of the final order, and by the coordination with the recovery
sector. At the beginning of the final phase, planners determine the final lot size based on the
number of primary products sold and the demand forecast. The final lot sizes reflect the
decision maker’s expectations on the failure rate of the part during the final phase and of the
demand for it as a spare part. Because the processes of product retum and parts recovery are
extremely uncertain, decision makers will not count on recovered parts yet. We assume that
they purpose to fulfill the entire spare-parts demand during the final phase by the final stock.
In future, this could be different when decision makers can use analogies or time series to
estimate more accurately, when and how many used products retum and in which condition.
Then the firm has to consider the retum rate of discarded primary products with recoverable
parts over time to take advantage of component recovery by reducing the size of the final
order.
Once the demand for a spare part is estimated, the final lot is manufactured with the actual
production equipment. Spare parts are stored in the stock of serviceable spare parts. Shortly
after the final order is fulfilled, the alternative to produce new spare parts is gone, for
example because suppliers discontinued the supply of electronic components, or the firm
discarded the production equipment. Thus, the spare parts must be redesigned if the spare-
parts demand during the final phase is underestimated.
Customers’ orders for spare parts accumulate in an order backlog until they are processed.
We modeled a target delivery delay to consider that orders have to be fulfilled within a
certain delivery time to meet customer requirements. In particular, the delivery of spare parts
of capital goods is often critical. For medical equipment or similar capital goods, the maximal
tolerable machine downtime is a day or two. The firm fills orders from the stock of
serviceable spare parts, initially from parts produced in the final lot. If the stock of
serviceable spare parts falls below a target value, the firm establishes a desired recovery rate,
which is passed on to the recovery sector of the model. Once a redesign of the spare part has
become necessary, and as soon as the firm can produce new spare parts, it must coordinate
the production process and the recovery process. The difference between the desired recovery
rate and the actual possible recovery rate determines the desired production rate. At the end
of the final phase, excess spare parts are sent to materials recycling.
OEMs use forecast procedures to determine spare- parts demand, because the lead-time to
manufacture, and to deliver spare parts to the customer is higher than the target delivery time.
It is assumed, that the producer uses first-order exponential smoothing, which is a realistic
model of the forecasting process of spare parts in many firms (Ihde, Merkel, and Hanning
1999).
2.3 Reverse logistics and recovery sector
This sector models how the recovery company takes back discarded products, stores them,
disassembles them to recover spare parts, and passes them on for materials recycling. The
demand for recovered parts from the spare-parts supply sector is met by depleting the stock
of recoverable spare parts. Figure 2.3 depicts an influence diagram, which shows the
relationships between the main variables of the spare- parts supply and recovery processes.
The starting point of the diagram is the variable “sales of equipment”. Minus signs alongside
links mean that the dependent variable decreases when the independent variable increases; for
example, the higher the “delivery rate” the lower is the “order backlog”. Positive signs
alongside links means that the dependent variable increases when the independent variable
increases; for example, the higher the “production rate“, the higher is the “stock of
serviceable spare parts’. We marked an arrow with double lines, whenever changes of
independent variables give effect to dependent variables with considerable delay; for
example, the “redesign of the spare part” requires six months.
Disassembling discarded equipment to gain recoverable parts fills the stock of serviceable
spare parts. Residuals of the disassembled equipment and used retumed equipment, from
which the recovery company does not recover parts, are sent to materials recycling. One
should consider that it is not possible to recover parts from all of the retumed equipment,
because of wear and tear. The aging of the equipment in use tends to result in a higher share
of nonrecoverable parts. Thus, we modeled age-dependent recoverability yields for obsolete
equipment. Recovery companies passes on all units of equipment to material recycling if
there is no demand of recoverable parts. At the end of the final phase, excess recoverable
parts are also sent to materials recycling.
The recovery company obtains its stock of discarded equipment from the discarded
products retumed out of the stock of obsolete devices held by former users. We modeled two
streams of equipment retums. The waste-stream retum rate contains the discarded units of
equipment that are accepted passively. In comparison, if the firm needs used recoverable
parts, it will acquire only that equipment which contains recoverable parts. Hence, it could be
expected that the recovery yield rate of that equipment, which is retumed market driven, is
higher than the units of equipment that are retumed waste driven (Guide, and van
Wassenhove 2000). The success of the product acquisition policy depends highly on how
customer friendly the take- back systems are, or how much the CLSC is willing to pay for a
unit of discarded equipment. The more the CLSC is engaged in the take back process of his
products, the lower is the quota of obsolete equipment which former users sent to alternative
disposal options, for example independent recycling companies. Additionally, it is assumed
that it takes more time and more effort to acquire a unit of equipment with recoverable parts
the lower the proportion of that equipment to the number of equipment in current use plus the
number of units of equipment in the virtual stock of obsolete equipment. The measures to
acquire obsolete units of equipment increase the costs for reverse logistics per retumed
equipment.
delivery4—— order
delay backlgg
order rate
delivery
rate F by ? +
+ lecision to .
redesign the spare desired
part recovery rate Y
difference between
on desired recovery rate and \ demand for
d actual recovery rate spare’ gaits
redesign of the
spare part
stock of stock of
serviceable spare + recoverable parts
arts ability to produce
new spare parts
+
\ 4 n
8
desired
production recovery prodyction rate
rate. rate
sales of
equipment
Figure 2.3: Influence diagram of the spare-parts supply and parts recovery processes
The causes and effects, which take place in the reverse logistics and recovery sector, are
depicted in the following influence diagram (figure 2.4).
alternative
disposal rate + aging of proportion of obsolete
equipment’ > —SCS«uipmentwith
sales of recoverable parts
equipment *
stock of obsolete
products effort to aquire obsolete
products with
recoverable parts desired equipment:
returns with
recoverable parts
return rate +
waste stream
return rate
market driven
stock of retumed equipment
"
materials
recycling rate of
7 equipment wit
recoverable parts =| ‘disassembly
Lo rate
desired , ‘tock of
disassembly rate recoverable parts
Figure 2.4: Influence diagram of the reverse logistics and recovery sector
2.4 Evaluation sector
Finally, the evaluation sector of the model tracks the costs for production, for reverse
logistics, for obsolete equipment acquisition, for disassembling discarded equipment, for
recovering components, for material recycling, for storing new and recoverable parts and for
redesigning spare parts. With these costs and the revenues from selling spare parts, we can
determine the net present value of the simulated policy. Therefore, an adequate target rate
must be determined outside the model. In the following, we assume that the costs of the
redesign of the spare part and the production costs per part after the redesign are generally
higher than recovering parts from discarded equipment and acquiring obsolete equipment
from former users. This assumption agrees with the authors’ experiences in practice.
Additionally, it allows us to focus on the primary purpose of the paper, to design efficient
control systems in CLSCs with parts recovery and spare- parts supply.
3. WORKING EXAMPLE
3.1 Base Case
As a working example, we are envisioning an OEM of medical equipment who wants to
reorganize its management of spare parts that are in the final phase. The firm often had to
redesign spare parts in the past, because its final stocks were not sufficient. Additionally, the
firm usually started the redesign of the spare part too late and so customers had to wait for
their spare parts. The firm operates a waste driven system to fulfill environmental law
requirements, but it is not engaged n using recovered parts from discarded equipment as
spare parts until now. We applied the model to a typical product of that firm. The firm
produced 3666 pieces of that equipment during 30 months. After it stopped selling that
product, it must guarantee a supply of spare parts for nine years (138 months). The target
delay between the order of a customer and delivery is two days. We assume that when the
primary product fails, repairing it requires one piece of a specific spare part. The entire spare-
parts cemand in the final phase amounts to 3,171 parts. The age-dependent obsolescence
parameters go up constantly from 0.05 per cent to 67 per cent. The age-dependent repair
parameters cut down continually from 98 per cent to 18 per cent.
To focus on the system behavior in the final phase, the initial inventory of the serviceable
spare part matches the spare-parts demand that occurs until the beginning of the final phase.
The initial parts production capacity is 300 parts per month until the final lot is produced. If a
redesign of the spare part becomes necessary, the assumed production capacity is set to 100
parts per month. In average, the OEM needs one month to become aware of a spare part out-
of-stock situation. Then, it requires six months to redesign the spare part. To study the
dynamic behavior of the spare-parts- supply system when the firm underestimated spare- parts
demand, we cut the optimal final lot size (the entire spare-parts demand during the final
phase) by about 20 per cent. Figure 3.1 depicts the dynamic behavior of the main flows. The
firm starts to produce the final order after the 21% month, to be able to discontinue the
production of the part along with the cessation of the equipment. After the 100" month, the
final stock is depleted and spare parts orders accumulate in the order backlog. Since the
service phase lasts until the 138" month, the firm has to redesign the spare part, which lasts
12 months. Customers have to wait several months before they can use their equipment again,
and the firm has to indemnify the customers, for instance by making another piece of
equipment with similar functions available to the customer. Costs accrue through the
measures of compensation and the redesign of the spare part and reduce the net present value
considerably. Since the firm does not recover parts, it passes on discarded equipment directly
to material recycling. The firm’s waste driven system achieves a return rate of 20 per cent of
the obsolete equipment that is former users pass on 80 per cent of their obsolete equipment to
alternative disposal options.
ES A
PR 1 pr
# [ah
og) /
‘ er ai a!
B ed t hee cueberr
3] By bg} rattan 2
=.= +] Material Recycling Rate (MR) = = = -[ Production Rate (PR)
EquipmentiMonth Parts/Month
Equipment Sales (ES) sonia] Order Rate (OR)
Parts/Month Parts/Month
Figure 3.1: Dynamic behaviour of the main flows (base-case)
The results of the base-case simulation run show that the spare parts control system of the
firm is inadequately designed. A fter the firm produced the final lot, there is no other option to
response to shortages, than to redesign the spare part. Since the final stock is inadequately
monitored, the redesign starts too late to avoid high order backlogs for a long time. In the
next sections, we explore modifications of the control system and include the option to
recover parts from discarded equipment.
The primary purpose of the study is to raise the robustness of the management control
system, and thereby to improve the performance of the CLSC. The system is assumed robust,
if an underestimation of demand does not necessarily result in a redesign of the spare part,
which is the worst case. Thus, we investigate strategies with regard to the maximum
allowable error in forecasting the entire spare- parts demand.
Considering used parts as an additional option for managing spare parts in the final phase
the firm has to incorporate the fact that during the final phase the maximum achievable
number of recoverable parts decreases permanently. The age-dependent recoverability yields
diminish from 92 per cent to 17 per cent for obsolete equipment.
Furthermore, it is assumed that only 20 per cent of the obsolete units of equipment are
retumed to the CLSC, except the firm acquires obsolete units of equipment in an active way.
3.2 Non-robust policies
In-depth simulation studies show that reactive policies are not robust. Reactive policies are
characterized by the OEMs behavior not to react to shortages in advance. The OEM gains
recoverable parts from its waste-stream system not until orders accumulate in the order
backlog. Even the OEMs approach to compensate shortages by acquiring obsolete units of
equipment with recoverable parts actively to increase the product returns of the waste- stream
system is not successful. A better performance of the policy is restricted by the late start to
acquire units of obsolete equipment if the final stock size is not enough. Because it requires
too much time to acquire and retum the units of equipment with recoverable parts, orders
begin already to cumulate in the backlog and cause the delivery delay to rise.
3.3 Systemwide-inventory policy
What is required is to initiate the acquisition of units of equipment with recoverable parts
early enough, so that the needed parts are in place by the time the final order stock is
depleted. Additionally the redesign of the spare part has to be set off early enough to avoid
that the delivery delay rises above the target delivery delay. To cope with those requirements,
it seems to be necessary to implement a forecasting procedure within the control management
system. Coyle defines “forecasting as the act of giving advance waming in sufficient time for
beneficial action to be taken.” ((Coyle 1978) pp. 237) The variable that has to be forecasted is
the systemwide-inventory of parts. The systemwide parts inventory equals the stock of
serviceable spare parts minus the order backlog. We integrate forecast procedures into the
model that are based on double exponential smoothing, one of the most common methods
used in spare-parts management (Ihde, Merkel, and Hanning 1999).
Whenever the systemwide- inventory is forecasted to be depleted, the CLSC has to start to
acquire obsolete equipment or to redesign the spare part. Since the recovery of parts is
preferred in comparison to the spare part redesign, the forecast horizon of the forecast
procedure, which determines the beginning of the acquisition of obsolete units of equipment,
has to be sufficiently longer than the forecast horizon of the forecast procedure, which
determines the redesign of the spare part. To determine the redesign of the spare part, a
forecast horizon of six months, the time needed to redesign the spare part, has been chosen.
The forecast horizon to determine the acquisition of dbsolete equipment has been set to
twelve months. As soon as the CLSC begins to actively collect units of obsolete equipment,
the computation of the systemwide- inventory is modified. The systemwide- inventory of parts
then equals the inventory of serviceable spare parts plus the inventory of recoverable parts
plus the actual recovery yield rate multiplied with the stock of retumed equipment minus the
order backlog. Thereby, the systemwide- inventory will be larger, and the probability that a
redesign will be necessary is lower.
f RR
TR
[i] RTR
BOR
MR
Bog ot
7 i
' 1 t.
1 1 mh,
vd ' Fs )*.
1 1 f
\ /
pod 1 —}—3 Ea)
1 q iF \
tft
1 ' id
1 Woof
109 pe
rT
) al
1 o
a
‘i |
Let” .
a etl wut
om: ba bg
‘ime (Month)|
=. — | -Material Recycling Rate (MR) == —[ Production Rate (PR) snus men Order Rate (OR)
CaupertMox Pareto Porson
— —|-Recoverable Take-Back Rate (RTR) +4 Total Take-Back Rate (TTR) —...}Recovery Rate (RR)
eupment Monk Fauperen Worth Peremone
Figure 3.2: Simulation run of the systemwide-inventory policy
The systemwide-inventory policy is more robust in comparison to the reactive policies we
have mentioned before, because it allows an underestimation of the spare-parts demand
during the final phase by about four per cent.
Figure 3.2 depicts the dynamic behavior of the main flows of the systemwide- inventory
policy within a simulation run where the entire spare-parts demand in the final phase has
been underestimated by about four per cent. The CLSC forecasts an out of stock situation at
the 110th month. Then, it begins to acquire obsolete units of equipment and the material
recycling of recoverable parts is stopped. Since spare-parts demand forecast errors amount
often up to 10 per cent and more, the results of the systemwide- inventory policy are not
sufficient.
3.4 Early-warning-system policy
Another way, which seems to be promising, is to combine the systemwide- inventory policy
with an early waming system. Therefore, decision makers, who determine final orders, have
to set a schedule how they believe the systemwide- inventory will change over time during the
final phase. During the final phase decision makers have to compare the planned schedule of
the systemwide- inventory with the actual schedule of the systemwide- inventory continually.
Thereby, it will be possible to start to acquire obsolete equipment as soon as there is an
obvious risk, that the final stock will not be sufficient to fulfill the entire spare- parts demand
during the final phase. The firm needs one month to initiate measures to close the gap
between the planned systemwide- inventory and the actual systemwide- inventory.
10
Px
ime (Mont
cece croee: (Qptimal systemwide invenibry (Phrts)]
—————_ Planned'systemwide inventory (Parts)
Figure 3.3: Planned and optimal systemwide-inventory within a simulation run of the early-warning- system
policy
Figure 3.3 shows the different behavior between a planned schedule of the systemwide-
inventory when decision makers underestimate spare-parts demand by about 24 per cent and
the optimal schedule of the systemwide- inventory.
Figure 3.4 depicts the result of the simulation run with the given schedule. Even an
underestimation of 24 per cent does not result in a spare parts redesign. The main reason is
that there is already a difference between the planned systemwide- inventory and the actual
systemwide-inventory near the beginning of the final phase. Therefore, the CLSC starts to
acquire obsolete equipment and stop the material recycling of units of equipment with
recoverable parts very early. Thus, in this case, the early- warning-system policy is very
robust.
The use of the early- warning-system policy is the more robust the more the decision
maker underestimates spare-parts demand at the beginning of the final phase, because it
causes the firm to acquires more units of equipments with recoverable parts.
1
TTR
Ed rtR
OR
MR
ah
ra bi
1 :
+! f Ea
1
q 1 Fa
qo g
' 1 |
yoy
yoy
ie [c)
| 4 wor Ts
le yor ®,
bean
a KT
Fh ace al aes i om el
[i] & iz] i] fiz] 6d
ime (Monthy
—- — | Material Recycling Rate (MR) Production Rate (PR) cones ef Order Rate (OR)
Equipment Month ParsMonth Pars/Month
— —|.Recoverable Take-Back Rate (RTR Total Take Back Rate (TTR) L.. LRecovery Rate (RR)
Equipment/Month Equipment/Month Parts/Month
Figure 3.4: Simulation run of the early-warning- system policy
4. CONCLUSIONS
We presented the development of a system dynamics model for designing control
management systems within CLSCs with parts recovery and spare-parts supply. The model
allows decision makers to study the cause and effects, which takes places in a closed-loop
supply chain. We take into account different forms of product take back (active product
acquisition and waste stream), and different policies such as, when to stop sending
recoverable parts to materials recycling, when to acquire units of equipment with recoverable
parts and when to begin to redesign a spare part. Considering the results of section 3.4, it
seems to be necessary to strengthen the monitoring of final stocks as well as the stocks of
recoverable parts within closed-loop supply chains. Therefore, it is necessary to improve the
information exchange between producers of EEE, who have to provide spare parts and
recovery companies that are involved in the take-back and recovery process of the producer’ s
goods (Spengler, and Stélting 2003).
5. ACKNOWLEDGEMENTS
This article is based on research work within the project “StreaM” (substance flow oriented
closed-loop supply chain management in the electrical and electronic equipment industry),
which is kindly sponsored by the German Ministry of Education and Research under the label
01RU0036. We thank the sponsor for its support in the name of all participants.
6. REFERENCES
Coyle, R. G. 1978. Management system dynamics. Chichester: Wiley.
Fleischmann, Moritz, Nunen, Jo van, and Ben Graeve. 2002. Integrating closed-loop supply chains and spare
parts management at IBM. Rotterdam: ERIM.
Fleischmann, Moritz. "Reverse logistics networks structures and design." Business Aspects of Closed-Loop
Supply Chains. Ed. V. Daniel, R. Guide, and Van Wassenhove, Luk N. Pittsburg (USA): Camegie Mellon
12
University Press, 2003.
Fleischmann, Moritz. 2001. Quantitative models for reverse logistics. Berlin: Springer.
Fortuin, Leonard. 1980. The all-time requirements of spare parts for service after sales - theoretical analysis and
practical results. International journal of operations and production management 1 (1):59-69.
Guide, V.D.R. Jr., and L.N. van Wassenhove. 2000. Managing product returns for remanufacturing. Production
and Operations Management 10 (2):142-155.
Hagen, M., Iding, M., and H. Sudhoff. "Losungsansatze zur Sicherstellung der Nachserienversorgung in der
Automobilelektronik." Elektronik im Kraftfahrzeug. Ed. Gesellschaft Fahrzeug- und Verkehrstechnik. VDI
Berichte. Diisseldorf: VDI-Verl., 2000. 953-978.
Hesselbach, Jiirgen, Mansour, Markus, and René Graf. "Erweitertes Supply Chain Management durch
ganzheitliches Ersatzteilmanagement." Logistikplanung und -management: 8. Magdeburger Logistik-Tagung,
Magdeburg, 14./15. November 2002. Ed. M. Schenk, and Ziems, D., Inderfurth, K. . 2002. 225-246.
Thde, Gésta B., Merkel, Helmut, and Ralf Hanning. 1999. Ersatzteillogistik: theoretische Grundlagen und
praktische Handhabung. Miinchen: Huss.
Schroter, Marcus, and Thomas Spengler. 2004. Designing control management systems for parts recovery and
spare parts management in the final phase within closed-loop supply chains. International journal of integrated
supply management 1 (2):158-179.
Schulz, Ekkehard. 1977. Grundlagen zur Planung von Ersatzteilfertigungen. Mainz: Krausskopf.
Spengler, Thomas, and Marcus Schroter. 2003. Strategic management of spare parts in closed-loop supply
chains - a system dynamics approach. Interfaces 33 (6).
Spengler, Thomas, and Wiebke Stélting. "Recycling-oriented information management in closed loop supply
chains in the electrical and electronic equipment industry." Strategy and organization in supply chains. Ed. S.
Seuring, M. Miiller, M. Goldbach, and Schneidewind, U. Heidelberg: Physica-Verl., 2003. 353-368.
Sterman, John D. 2000. Business dynamics: systems thinking and modeling for a complex world. Boston:
Irwin/McGraw- Hill.
Teunter, Ruud H., and Leonard Fortuin. 1999. End-of-life service. International journal of production
economics 59 (1-3):487-498.
ZVEI. 2002. Langzeitversorgung der Automobilindustrie mit elektronischen Baugruppen. Frankfurt am Main.
Zamudio - Ramirez, Pavel. 1996. The Economics of Automobile Recycling. Cambridge.
13
