Int. J.

Production Economics 133 (2011) 233–242

Contents lists available at ScienceDirect

Int. J. Production Economics

journal homepage: www.elsevier.com/locate/ijpe

A minimum cost flow model for level of repair analysis

R.J.I. Basten a,, M.C. van der Heijden b, J.M.J. Schutten b
a
Eindhoven University of Technology, School of Industrial Engineering, Department of Operations, Planning, Accounting, and Control, P.O. Box 513, 5600 MB, Eindhoven,
The Netherlands
b
The University of Twente, School of Management and Governance, Department of Operational Methods for Production and Logistics, P.O. Box 217, 7500AE, Enschede,
The Netherlands

a r t i c l e in fo abstract

Available online 9 April 2010 Given a product design and a repair network for capital goods, a level of repair analysis determines for
Keywords: each component in the product (1) whether it should be discarded or repaired upon failure and (2) at
Maintenance which location in the repair network to do this. In this paper, we show how the problem can be
Supply chain management modelled as a minimum cost flow problem with side constraints. Advantages are that (1) solving our
Level of repair analysis model requires less computational effort than solving existing models and (2) we achieve a high model
Mixed integer programming flexibility, i.e., many practical extensions can be added. Furthermore, we analyse the added value of
Minimum cost flows modelling the exact structure of the repair network, instead of aggregating all data per echelon as is
common in the literature. We show that in some cases, cost savings of over 7% can be achieved. We also
show when it is sufficient to model the repair network by echelons only, which requires less input data.
& 2010 Elsevier B.V. All rights reserved.

1. Problem setting and literature
For capital goods, such as military naval equipment, MRI-scanners,
or trains, customers increasingly take total life cycle costs (LCC)
into account in their purchasing decisions (Ferrin and Plank,
2002). Also, we see a trend that customers outsource activities for
system upkeep to the original equipment manufacturer (OEM)
using service contracts that guarantee a certain service level
against fixed yearly costs. For the OEM, this can be attractive, since
selling services is generally more profitable than selling products
(Deloitte, 2006; Murthy et al., 2004; Oliva and Kallenberg, 2003).
This means that it becomes important for the OEM to take the costs
of maintenance into account when designing new products. Costs
of more reliable components can be earned back by lower
maintenance costs during the product life cycle.
Generally, capital goods are repaired by replacement, which
means that a failed component is removed from the system and
replaced by a functioning one. A defective component can either
be discarded or repaired. If it is discarded, a new component
needs to be purchased. If the component is repaired, this often
means that a subcomponent failed and is replaced by a
functioning one. The subcomponent should in turn be repaired
or discarded itself. The system is thus seen as a multi-indenture
system such as shown in Fig. 1.
 Corresponding author. Tel.: + 41 40 247 3947.
E-mail addresses: r.j.i.basten@tue.nl (R.J.I. Basten),
m.c.vanderheijden@utwente.nl (M.C. van der Heijden),
j.m.j.schutten@utwente.nl (J.M.J. Schutten).
If a component should be repaired, it should also be decided
where to do that. For example, if we consider military naval
equipment, repairs can be performed on board the ship, at its
marine base, a central depot, or even at the OEM. A network that
connects all ships, bases, depots, and the OEM is called a multi-
echelon repair network, see Fig. 2 for an example.
In an early stage of the product life cycle, decisions upon
discard or repair, and location of maintenance activities should be
taken. These decisions are covered by a level of repair analysis
(LORA). A LORA model should decide for a given product design and
repair network:
 which components to repair upon failure, and which to
discard,
 for each component that will be repaired, where in the repair
network to do this, and,
 for each required resource (e.g. test or repair equipment),
where in the network to install it.
such that the lowest possible life cycle costs are achieved. Those
costs consist of both costs that are variable in the number of
failures and fixed costs. Variable costs include costs of hiring
service engineers and transportation of components; fixed costs
include costs for resources such as test equipment and tools.
The availability of the installed base is not considered in the
LORA. Instead, the spare parts stocking problem is solved using the
decisions that result from the LORA as an input (e.g., using METRIC-
like methods, see Sherbrooke, 2004; Muckstadt, 2005). The goal in
this problem is to allocate inventory in a repair network such that

0925-5273/$ - see front matter & 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijpe.2010.03.025
234 R.J.I. Basten et al. / Int. J. Production Economics 133 (2011) 233–242
Fig. 1. System structure.
Fig. 2. Balanced repair network.
a certain availability of the installed base is achieved against the
lowest possible spare parts costs.
A lot of literature and commercial tools exist for the spare
parts stocking problem, but surprisingly, the scientific literature
on LORA is limited, as we show below. Commercial tools exist, for
example PRICE HL (2007) and EDCAS (2009), but it is not clear how
they function.1 In practice, the LORA problem is often not solved
explicitly using a formal model, but implicitly using expert
knowledge and the decisions made for older products. Spread-
sheets are used to calculate the costs for a few scenarios only.
After solving the LORA problem and spare parts stocking problem, it
can turn out that spare parts costs for some components are very
high. In that case, a second iteration is sometimes made in which
another LORA decision is taken for those components. Obviously,
it is not guaranteed that the optimal solution is found in this
sequential or iterative manner. Besides, performing such calcula-
tions is time consuming, and mistakes are easily made.
In the LORA, high resource costs typically causes that repairs
are performed upstream in the repair network, whereas high
transportation costs causes downstream repairs, close to the
installed base. So, components requiring no or cheap resources
should generally be repaired locally and components requiring
expensive resources should be repaired centrally. In a high tech
environment (like Thales Nederland), resources tend to be
expensive and thus repairs of components requiring resources
are generally planned at the central level. However, this leads to
high lead times and thus to high spare parts costs. Saranga and
Dinesh Kumar (2006) include the spare parts holding costs in the
LORA, but this can only be done in a simple way that is not
consistent with the actual spare parts stocking problem in
practice. The reason is that the LORA needs spare part costs
estimates on an individual component level, whereas it is known
that the actual costs result from a multi-item, multi-echelon
system approach like VARI-METRIC (see, e.g., Sherbrooke, 2004;
Muckstadt, 2005). Still, including spare part holding costs will
cause more downstream repairs in the service supply chain. In a
broader sense, a good LORA model is also useful because it could be
used as a building block in an iterative approach in which the LORA
and the VARI-METRIC spare parts optimization are solved iteratively
(this is ongoing further research at the time of writing).
1
We noticed this ourselves, but it is also noticed by experts who have been
using these tools, and in the literature (e.g., Brick and Uchoa, 2009).
Fig. 3. Aggregated repair network.
The LORA problem itself is nontrivial due to:
 the multi-indenture system structure: decisions for a sub-
component are only needed if the parent component is
repaired;
 the multi-echelon repair network: if a component is repaired
at a certain location, it is illogical to repair a subcomponent at
a location downstream in the network; and
 the shared resources: some components share resources that
are needed to repair them. This means that it is often cost
effective to repair these components at the same location.
As mentioned, the scientific literature on LORA is limited. Barros
(1998), Barros and Riley (2001), Saranga and Dinesh Kumar
(2006) and Basten et al. (2009) all assume infinite capacity of
resources and aggregate all data per echelon. This means that
every three echelon repair network would be represented as in
Fig. 3. The key differences between these papers are:
 Barros (1998) and Barros and Riley (2001) assume that all
components at one indenture level share the same resource;
 Saranga and Dinesh Kumar (2006) assume that components do
not share resources and every component needs exactly one
resource;
 Basten et al. (2009) relax the assumptions made in the
forementioned papers in that sets of components sharing
resources can be defined freely.
Brick and Uchoa (2009) do not aggregate all data per echelon
level, but they model only 1 echelon and effectively assume 2
indenture levels. Integrated in their LORA is the decision of which
facilities to open.
Alfredsson (1997) presents a model for the integrated problem
of LORA and spare parts stocking. Since he combines the two
problems, he has to make some simplifying assumptions. He
assumes one indenture level only, and two echelons levels, but
the extension to more echelon levels is straightforward. Further-
more, he assumes that each component has its own tester
(resource) and one multi-tester exists. This multi-tester can be
used for one component and adapters can be added in a fixed
order. The adapters enable the multi-tester to be used for the
repair of additional components.
In this paper, we model the LORA problem as a minimum cost
flow model with side constraints. We can solve problem instances
much faster using our new formulation than using the formula-
tion of Basten et al. (2009). We compare our model with theirs,
because it is the most generic model that can cope with multi-
indenture systems and multi-echelon repair networks. Further-
more, our formulation allows for many extensions in an elegant
manner, including repair probabilities, no-fault-found probabil-
ities, and equipment with finite capacity. See Basten et al. (2010),
in which a case study is solved at Thales Nederland, a
manufacturer of naval sensors and naval command and control
systems. In the current paper we analyse modelling the exact
structure of the repair network instead of aggregating all
information per echelon, which is not possible with the existing
 R.J.I. Basten et al. / Int. J. Production Economics 133 (2011) 233–242
Fig. 4. Unbalanced repair network.
multi-echelon multi-indenture models. This means that we can
differentiate between the repair networks in Figs. 2 and 4, which
may lead to different decisions and costs. For example, for a
European OEM, we may decide to repair certain components at the
OEM for the installed base in Europe, whereas we repair the same
components for the installed base in Asia at local repair shops.
A drawback is that we need more information of the structure
of the repair network and the assignment of the installed base to
the various locations. This information can be hard to obtain early
in the product life cycle when a LORA is performed. In this paper,
we analyse the impact of modelling the exact repair network in
terms of reduction of life cycle costs. We find that in repair
networks that are unbalanced in the locations, such as shown in
Fig. 4, significant cost reductions of sometimes more than 7% can
be achieved. We also find that in repair networks that are
balanced in the locations, no significant cost reductions can be
achieved by modelling the exact network, even if costs differ
across the repair network. In those cases, it is not worth the
additional effort to collect input data.
Summarizing, the added value of this paper to the literature is
three-fold:
 we provide a new minimum cost flow model for the LORA
problem that can be solved faster than existing models;
 the new model formulation allows for many practically
relevant extensions;
 we provide insight in the impact of modelling the exact repair
network on the life cycle costs and show under which
conditions data aggregation per echelon is sufficient.
The structure of the remainder of this paper is as follows.
Section 2 explains which inputs a LORA needs and what assump-
tions we make. Section 3 models the LORA problem as a minimum
cost flow model with side constraints. Section 4 gives the results
of our experiments. The paper ends with conclusions and
directions for further research in Section 5.
2. Model assumptions and input data
We model a multi-indenture system structure, with F being the
set of all components. The subcomponents or children of a (father)
component f A F form the set Gf . For example, in Fig. 1: G1 ¼ f2,3g.
For each component, there are three possible decisions d A D at
every location: discard, repair and move. Discard means that the
component is scrapped and a new one should be bought. Under the
repair option, we assume that any repair is successful and that a
component on which a repair is performed, has really failed
(in practice, a ‘no-fault-found’ may occur), although both assump-
tions can be generalized (see Basten et al., 2010). The move option
means that the component is moved to the location at the next
higher echelon for further decision making. Note that the ‘move’
option does not exist at the highest echelon level, the central depot.
A system in the installed base is always repaired by replacing a
subsystem. Therefore, the first indenture level that we model is
235
that of the subsystems (also called line replaceable units or LRUs):
Component 1 in Fig. 1. These components form the set Fs. A failure
in component f is due to a failure in subcomponent g with
probability qfg. In that case, repairing component f means
replacing subcomponent g. For subcomponent g, we have to
decide whether to discard, repair or move it. It is possible that two
(or more) subcomponents fail at the same time, which could
P
mean that for some component f: g A Gf qfg 4 1. There is also a
probability qff that a failure can be repaired directly, without
replacing a subcomponent. No further decisions have to be taken
if a component is repaired directly.
We model a multi-echelon repair network, with L being the set
of all locations. The structure is divergent, that is, each location in
the network has a single upstream location to which it can move
its failed components. All locations that supply failed components
to location l form the set Fl . For example, in Fig. 2: F1 ¼ f2,3g. All
the system locations (echelon 1), so the locations l that have
Fl ¼ |, form the set Ls.
If a system fails, this is due to a failure in a subsystem, so
components f A F s . We define mfl as the average yearly number of
failures of component f A F s at location l A Ls . For each failure,
decisions have to be taken as described above. If a certain decision
d A D is chosen for component f A F at location l A L, then variable
costs of cvfld are made each time this component fails.
We model resources r A R with Yr ¼ fðf ,dÞ j resource r
is required for decision d for failures in component f g. For exam-
ple, Yr1 ¼ fðf1 ,repairÞ,ðf1 ,discardÞ,ðf2 ,repairÞg. We assume that the
resource capacity is infinite, although we can extend our model to
finite capacities (see Basten et al., 2010). Without loss of
generality, we have chosen to minimize the average total costs
per year with our definition of mfl. Therefore, we define fixed
costs cfrl per year if we decide to use resource r at location l. These
costs are, for example, the yearly depreciation costs of the
resource and costs of capital.
3. Minimum cost flow model
This section explains how the LORA problem can be modelled as a
minimum cost flow model with side constraints (see, for example,
Ahuja et al., 1993 for an overview of minimum cost flow models).
To define the graph underlying the flow model, we need four
different node types: Source nodes ðv A V s Þ are used to represent the
occurrence of failures in subsystems (indenture 1) at system
locations (echelon 1). The flows from these source nodes arrive at
decision nodes ðv A V d Þ where a decision is made between the three
available options: discard, repair and move. The variable costs are
attached to the outgoing arcs of the decision node, each represent-
ing an option. If repair is chosen, then a transformation node
ðv A V g Þ is used to represent that a failure in a parent is due to a
failure in any of the children. If no decisions need to be made
anymore, the flow goes to a sink node ðvA V t Þ. We model the use of
resources by side constraints on the minimum cost flow model: If
the outgoing arc of a decision node represents an option for a
component that can only be chosen if a resource is available, then
the capacity of this arc is 0 if the resource is not available.
Section 3.1 explains how we construct the graph that forms
the basis of the flow model. Section 3.2 shows how the resources
are added to the model as side constraints. Section 3.3 provides
the formal model formulation.
3.1. Construction of the graph
In this section, we explain how the node types are used in the
model and we define the incoming and outgoing arcs (v,w) and
the relevant cost parameters. To illustrate our model, we use the
236 R.J.I. Basten et al. / Int. J. Production Economics 133 (2011) 233–242
Fig. 5. Example: System structure.
Fig. 6. Example: Repair network.
Fig. 7. Source node. (Failures of component 1, shown by the number next to the
arc, originate at the source node.)
following example throughout this section: We have a two-
indenture system with three components (G1 ¼ f2,3g, see Fig. 5)
and a two-echelon repair network with three locations (see
Fig. 6). We show an example of each node type, and after we have
introduced all the node types, we show the complete flow model.
In all figures related to this example, a number next to an arc
represents a component, whereas a letter next to an arc
represents an option (r¼repair, d ¼discard, m¼move).
3.1.1. Source nodes
Source nodes represent occurrences of failures of a certain
subsystem at a certain system location. In our example only
failures in component 1 occur at locations 2 and 3. This means
that we have two source nodes in our flow model, of which the
one for location 2 is shown in Fig. 7. The flow out of this source
node is equal to the yearly number of failures of component 1 at
location 2: m1,2. In general, for every component f A F s and every
location l A Ls , there is one source node vA V s with outflow
sv :¼ mfl .
3.1.2. Decision nodes
If a subsystem at a system location fails, there are three
options to choose from:
 move the component to the next higher echelon;
 repair the component, which means replacing a subcomponent
or repairing the component directly;
 discard the component.
In the flow model, it means that an arc originating at a source
node terminates at a decision node v A V d .2 Every arc going out of
the decision node represents one of the available options.
Fig. 8 shows the decision node for component 1 at location 2.
If arc (v,w) represents the repair option ‘r’, then the variable costs
for using this arc, cavw, are equal to the variable costs of option
‘repair’ for component 1 at location 2: cv1,2, repair. In general, the
variable costs, cavw, for using an arc ðv,wÞjvA V d are equal to cvfld,
where arc (v,w) represents decision d for component f at location
2
The source node could also be integrated in the decision node, but for clarity
we prefer to have a distinction between source and decision nodes.
Fig. 8. Decision node. (Failures of component 1 at location 2 go into the decision
node, represented as ‘1,2’ in the node. Options repair ‘r’, discard ‘d’, and move ‘m’
can be chosen.)
Fig. 9. Transformation node. (The flow going into the node results from a repair ‘r’
of component 1 at location 2, shown as ‘1,2,r’ in the node. Components 3 and 2
flow out of the node, respectively.)
l. In this way, the variable costs of the LORA model are attached to
the arcs originating at the decision nodes; all other arcs have zero
costs associated to them.
If the decision is taken to move component 1 from location 2 to
location 1, a decision has to be taken at location 1. This means that
the arc representing the move option ‘m’ in our example (Fig. 8)
terminates at the decision node representing component 1 at
location 1. In general, the arc representing the move decision for
component f at location k terminates at the decision node
representing component f at location l, where k A Fl ðF1 ¼ f2,3gÞ.
Note that at a node representing a component at the highest
echelon, location 1 in our example, the ‘move’ option is not
available. The arcs representing the repair and discard options are
discussed below.
3.1.3. Transformation nodes
Transformation nodes represent the repair of a parent
component. A flow representing failures in a parent component
goes into the node and is split according to the probability that a
failure in that parent component is due to each specific child
component. Each outgoing arc terminates at a decision node for a
child component. In our example, a failure in component 1 can be
caused by a failure in component 2 or 3 ðG1 ¼ f2,3gÞ. The arc
representing the repair decision ‘r’ for component 1 at location 2
(in Fig. 8) terminates at the transformation node that is shown in
Fig. 9. The two arcs originating at the transformation node
represent failures of components 3 and 2, respectively. Suppose
that 50% of those failures are caused by failures in component 2
and 40% are caused by failures in component 3, which means that
10% of the failures is repaired directly ðq1,2 ¼ 0:5, q1,3 ¼
0:4, q1,1 ¼ 0:1Þ. If the outgoing arcs (v,w) and (v,u) represent
components 3 and 2, respectively, then gvw ¼0.4 and gvu ¼0.5.
Since no further decisions need to be taken for the failures that are
repaired directly, it is not necessary to model them.
In general, transformation nodes vA V g represent the repair of
a parent component f by replacement of any of the subcompo-
nents g A Gf at location l. If arcs (v,w) represent components
g A Gf , then the factor of inflow in node v that flows out on arc
(v,w), gvw, is equal to qfg.
 R.J.I. Basten et al. / Int. J. Production Economics 133 (2011) 233–242 237

3.1.4. Sink nodes at location 1, 2, 3, respectively. If flow resource 3 is available,

If no other decisions need to be taken after a certain decision then the arc representing repair of component 1 at location
has been made, then the flow goes to a sink node. In the example, 3 is enabled. This means that O3 ¼ fðv,wÞjarcðv,wÞ denotes
arcs that terminate at a sink node represent the decision ‘discard’ decision ‘repair’ for component 1 at location 3g. The costs of
for any component at any location and the decision ‘repair’ for enabling flow resource 3 (c3) are the yearly costs for resource 2
component 2 or 3 at any location. Fig. 10 shows a sink node. at location 3: cf2,3.
In general, for each location l and every resource r there will be
a flow resource e in the model such that Oe ¼ fðv,wÞjðv,wÞ
3.1.5. Example
denotes decision d for component f at location l with ðf ,dÞ A Yr g
We already showed parts of the flow model that results from
and ce :¼ crlf . Note that fixed costs are only related to the arcs
the LORA problem that we used as an example throughout this
originating at the decision nodes.
section. Fig. 11 shows the complete resulting flow model. The
dotted arcs for component 3 should be replaced by flows similar
to the flows for component 2. We left them out to improve the 3.3. Flow model formulation
readability of the figure.
In our flow model, we have two types of decision variables:
3.2. Modelling resources as side constraints
 Xvw, the amount of flow through arc (v,w);
In some cases, resources r are needed before a certain decision  Ye, a binary variable that indicates whether flow resource e is
can be taken for certain components. For example: Resource r ¼2 bought.
may be needed if and only if component 1 is to be repaired.
In that case, Y2 ¼ fð1,repairÞg. At all three locations in our Let A be the set of all arcs and E the set of all flow resources.
example repair network, we can decide whether we want to The resulting model formulation is
buy resource 2. Therefore, we distinguish in the flow model three X X
a
so called ‘flow resources’ e¼ 1, 2, 3 that represent resource 2 minimize cvw  Xvw þ ce  Ye ð1Þ
ðv,wÞ A A eAE

subject to
Xvw ¼ sv 8ðv,wÞ A Ajv A V s ð2Þ
X X
Xuv ¼ Xvw 8v A V d ð3Þ
Fig. 10. Sink node. (The flow that is discarded ‘d’ goes into the sink.) ujðu,vÞ A A wjðv,wÞ A A

Fig. 11. Example: network flow problema. aNote: A number next to an arc indicates the component that is represented by that arc. A letter indicates the decision that is
represented by that arc: r=repair, d=discard and m=move.
238 R.J.I. Basten et al. / Int. J. Production Economics 133 (2011) 233–242

X
Xvw ¼ gvw  Xuv 8ðv,wÞ A Ajv A V g ð4Þ Table 1
ujðu,vÞ A A Input parameters.

Parameter (varied) Values

Xvw r M  Ye 8eA E, 8ðv,wÞ A Oe ð5Þ
# Components 500–1,000–2,000–5,000
Xvw Z 0 8ðv,wÞ A A ð6Þ # Echelons 2–3
# Indenture levels 2–3
Max. # resources per component 1–2
Ye A f0,1g 8E A e ð7Þ
Constraint (2) states that the outflow of each source node v is Parameter (not varied) Value
equal to sv and Constraint (3) assures that the inflow into any
# Resources 100
decision node is equal to the outflow. For any arc (v, w) going out
of a transformation node, Constraint (4) assures that the total
inflow into that transformation node is transformed to outflow on Parameter (not varied) Range
arc (v, w). Constraint (5) assures that only arcs are used that are
Yearly demand 0.05–5
enabled due to the availability of flow resources; they become Variable costs 50–1,000
uncapacitated. Arcs that are not in any Oe are not capacitated Fixed costs 500–10,000
either. For each arc (v, w) that is part of any Oe , we set the value of
the big M equal to the maximum possible value of Xvw.

Table 2
4. Computational experiments Comparison of optimization times (s).

# Components Basten et al. (2008a) Flow model

In Section 4.1 we compare the time it takes to solve problem
instances using the flow model and the model of Basten et al. 500 0.9 0.9
(2009). A more efficient optimization method is useful to solve 1,000 1.2 0.9
2,000 2.2 1.2
large problem instances, but it is also useful if a LORA is performed
5,000 20.6 4.2
multiple times. This is for example the case, if an iterative
procedure is used to solve a LORA and spare parts optimization
# Echelons Basten et al. (2008a) Flow model
simultaneously. Section 4.2 compares modelling the repair net-
work exactly and aggregating all inputs per echelon. We compare 2 2.5 1.2
the time it takes to solve problem instances, the cost reductions 3 10.0 2.4
that can be achieved by modelling the repair network exactly, and
the repair strategies that result from modelling the repair # Indenture levels Basten et al. (2008a) Flow model
network exactly and aggregating all data per echelon level.
For our tests, we generate instances of the LORA problem and 2 1.8 1.2
solve these using the CPLEX callable library version 11 (with default 3 10.7 2.4

settings), running under windows XP, service pack 2, on an Intel

Centrino Duo, 2 GHz with 2 GB ram. Although CPLEX 11 can use both Max. # resources per component Basten et al. (2008a) Flow model
cores of the dual core processor, it seldomly does for these
1 1.9 0.8
problems. 2 10.6 2.8

4.1. Comparison with the model of Basten et al. (2008a) in terms of

optimization time
We generate problem instances in which all data is aggregated
per echelon, corresponding to the assumptions of Basten et al.
(2009). In our model, this is equivalent to a network structure
with one location at each echelon. We use the problem instance
generator as described by Basten et al. (2009). We omit the details
here for sake of simplicity.
We vary the four input parameters that most heavily
influenced the optimization time in their paper: The number of
components, echelons and indenture levels, and the maximum
number of resources per component. If the maximum number of
resources per component is 2, this means that in order to repair a
component, at most 2 different resources are required. See Table 1
for the settings. For each combination of parameters, we
generated 25 problem instances. In total, this makes 4  2  2  2 
25 ¼ 800 test runs.
Table 2 presents the average time it takes to solve the LORA
problem instances for each parameter that we varied. It is clear
that our model increasingly outperforms the model of Basten
et al. if the number of components, indentures or echelons or the
maximum number of resources per component increases.
4.2. Effect of modelling the exact repair network
In this section, we show the circumstances under which
exactly modelling the repair network reduces the total costs and
those under which inputs can be aggregated. We also compare the
time it takes to solve these problems and the repair strategies that
result.
We generate instances that are realistic in practice, based on
our experience at Thales Nederland. We make comparisons (1) for
a base situation with a symmetrical repair network, and for an
asymmetrical network (2) in terms of the number of system
locations that is attached to each intermediate depot and (3) in
terms of the costs for moving and repairing components. Section
4.2.1 explains the problem instances we use and Section 4.2.2
discusses the results.
4.2.1. Problem instances
In all tests, the system structure consists of 25 components at
the first indenture (subsystems), 125 at the second level and 625
at the third level. We use random generators to construct 10
 R.J.I. Basten et al. / Int. J. Production Economics 133 (2011) 233–242 239

instances of a product structure with corresponding failure rates
and cost factors, see Appendix A for details.
For our tests, we use various repair networks, of which some
are balanced in terms of the number of system locations per
intermediate depot and some are not. We call them balanced and
unbalanced in the locations. The smallest balanced network
consists of one central depot, two intermediate depots and four
system locations (two per intermediate depot), see Fig. 2. We vary
the number of intermediate depots (2 or 10) and the number of
system locations per intermediate depot (2 or 10).
In the unbalanced networks, there are 2 system locations per
intermediate depot for half of the intermediate depots and 10
system locations per intermediate depot for the other half of
the intermediate depots. Below, we call the left half of the
intermediate depots with the attached system locations the ‘left
half’. The other intermediate depot(s) with the attached system
locations are called the ‘right half’. This holds for both balanced
and unbalanced networks.
Besides being unbalanced in the locations, repair networks can
be unbalanced in terms of the costs. In our tests, costs in the left half
and at the central depot are always equal. Repair and move costs in
the right half can differ from the costs in the left half and at the
central depot. We test what happens if the repair costs in the right
half are 0.5 or 2 times the repair costs in the left half (and at the
central depot). We say that the relative repair costs are 0.5 or 2. In
the same way, we test with relative move costs of 0.25 or 4. These
values are chosen because for a European OEM, the costs of moving
components to Asia can be a number of times as high as those costs
in Europe. Repair costs can differ as well, but the relative difference
is assumed to be smaller. Discard costs are assumed to be
approximately the same at all locations, since a main part of these
costs are due to the costs of buying a new component.
If we do not include spare parts costs in the LORA problem,
which is what happens usually in practice, we see that many
repairs are performed at a central location. The explanation is that
in that case only one resource of each type needs to be acquired. If
repairs are performed decentrally, many resources are needed. In
the problem instances that we discuss below, 86% of the costs for
resources would be made at the central depot if we do not include
spare parts costs. Integrating spare parts optimization into the
LORA is interesting future research, but it does not fit in the scope of
this paper. However, we add spare parts costs in a basic way, by
assuming lead times for all possible decisions (discard, repair and
move) and relating spare parts costs to these lead times. This is
elaborated on in more detail in Appendix A.
There are 10 or 25 resources (types of test equipment) and we
distinguish two cases for the number of resources that each
component requires. In the first case, 70% of the components
needs no resources, 20% needs one resource, and 10% needs 2
resources. To be short, we write ‘0.7–0.2–0.1’. The other case is
‘0.25–0.5–0.25’.
Summarizing, we have 7 experimental factors, namely the
number of intermediate depots, the number of system locations in
the left half, the number of system locations in the right half, the
relative move costs in the right half, the relative repair costs in the
right half, the total number of resources, and the number of
resources per component. This gives 2  2  2  3  3  2  2 ¼ 288
combinations. For each combination, we generate 10 problem
instances as described in Appendix A. In this way we prevent that
we draw conclusions based on one exceptional case.
4.2.2. Results
Table 3 presents for the 8 repair networks, the average time it
takes to solve the exact and aggregated model. It also shows the
average and maximal difference in optimal solution value for both
models. In each test we vary over all other parameters, as
explained in the previous section. We focus on the other
parameters below.
We see that the aggregated models can be solved much faster
than the exact models. Still, it took only 104 s to solve the most
time consuming problem instance using the exact model. This
means that the exact model can be solved fast enough to be used
in practice.
In networks that are balanced in the locations, the cost
reduction that can be achieved is limited: in our tests, it is 1.2%
at maximum and only 8% of the problem instances leads to a cost
difference of more than 0.5%. This limited cost reduction means
that it is hardly worthwhile to acquire the additional input data if
the repair network is balanced in the locations, even if the
network is unbalanced in the costs. However, in our tests, the
spare parts costs that are added to the variable costs, are always
balanced in the network, even if the remainder of the variable
costs for repair and move are unbalanced. In a global repair
network, the spare parts costs may be unbalanced too, due to
differences in lead times. If we vary the spare parts costs with the
move and repair costs in a network that is balanced in the
locations, we see cost differences between the exact and
aggregated model of over 5%. However, our basic way of
incorporating spare parts, does not allow us to analyse this in
detail.
If the network is unbalanced in the locations, interesting cost
reductions are achieved quite often: in our tests, the maximum
cost difference is 7.4% and for 29% of the problem instances, a cost

Table 3
Comparison of various repair networks.a

#Intermediate depots # System locations Optimization time Cost difference

per intermediate depot exact–aggregated

Left half Right half Exact (s) Aggregated (s) Mean (%) Max (%)

2 2 2 0.3 0.1 0.1 0.9

10 1.7 0.1 1.0 5.1
10 2 1.6 0.1 1.0 4.7
10 1.0 0.1 0.1 0.9

10 2 2 1.6 0.1 0.1 1.2

10 8.1 0.1 2.7 7.1
10 2 8.4 0.1 2.6 7.4
10 10.2 0.1 0.2 1.1

a
It may seem strange that there is a difference between row 2 and 3 (2 intermediate depots and unbalanced network). However, these cases are not the same because,
for example, if the relative repair or move costs are not 1, this affects more system locations in row 2 than in row 3.
240 R.J.I. Basten et al. / Int. J. Production Economics 133 (2011) 233–242

Table 4
Comparison of three most important parameters.

#Intermediate depots # Resources Cost difference

exact  aggregated

Total Per component Mean (%) Max (%)

2 10 0.7–0.2–0.1 1.1 3.9

0.25–0.5–0.25 0.1 1.2
25 0.7–0.2–0.1 0.5 1.9
0.25–0.5–0.25 2.4

10 10 0.7–0.2–0.1 2.0 6.1

0.25–0.5–0.25 2.1 6.9
25 0.7–0.2–0.1 2.5 6.8
0.25–0.5–0.25 4.0 7.4
Fig. 12. Resource costs in repair networks that are unbalanced in the locations.
(Number of intermediate depots, and algorithm that is used.)

Table 5
Example of different solutions.

Type of model Costs (  h1,000)

Variable Resources Total

Exact 43,925 12,404 56,329

Aggregated 42,241 15,308 57,549
Difference (%)  4.0 19.0 2.1

difference of more than 2.5% is achieved. We focus in more detail

on the networks that are unbalanced in the locations in Table 4, in
which we vary the three parameters that most heavily influence
the cost differences: the number of intermediate depots, the
number of resources, and the number of resources per
component. We see that high cost differences are mainly
Fig. 13. Total costs divided by number of system locations in balanced repair
achieved in problem instances with ten intermediate depots. If networks. ax, y (e.g., 2,10) means: x intermediate depots and y system locations per
there are also many resources in total and many components need intermediate depot.
one or two resources, a large cost reduction by modelling the
exact network is almost guaranteed. However, we cannot define a
broad category of problem instances in which we can guarantee
that there are no cost reductions possible.
If relatively small cost reductions can be achieved by
modelling the exact network, the solutions of the exact model
and the aggregated model can still differ substantially. If there are
multiple, really different solutions that lead to approximately the
same total costs, there can be other, more qualitative reasons to
choose for another solution than the one with the lowest total
costs. Industry might be interested in tools that can provide these
different solutions. To give an example, we focus on the problem
instances in which the repair network is unbalanced in the
locations, there are ten intermediate depots, the number of
resources per component is 0.25–0.5–0.25, and there are ten
resources. Table 5 presents that for these problem instances, Fig. 14. Resource costs divided by number of system locations in balanced repair
the costs of resources are 19% higher on average if the model is networks. ax, y (e.g., 2,10) means: x intermediate depots and y system locations per
aggregated. However, this is compensated for by lower variable intermediate depot.
costs, so that the total costs differ only 2.1%.
Fig. 12 shows that more in general, resource costs are higher if
we aggregate the data in an unbalanced network than if we model assumes there are six system locations per intermediate depot on
the exact network, especially if there are ten intermediate depots. average.
If, for those problem instances, we divide the amount of money Fig. 13 gives some insights into how costs (divided by the
that is spent on resources at the intermediate depots by the number of system locations) change if a network (that is balanced
number of system locations in the network, there is a clear in the locations) grows due to more system locations per
difference between the half of the network with two system intermediate depot (compare bars 1 and 2, and bars 3 and 4) or
locations per intermediate depot ðh153,000Þ and the half with ten more intermediate depots (compare bars 1 and 3, and bars 2 and
system locations ðh116,000Þ. Such a distinction cannot be made if 4). Notice that both bars 2 and 3 give results for networks with 20
the data is aggregated, because in that case, the model implicitly system locations, although the results differ substantially. The
 R.J.I. Basten et al. / Int. J. Production Economics 133 (2011) 233–242 241

reason is that repair at intermediate depot is much more
expensive if there are ten intermediate depots, than if there are
only two, due to the number of required resources. As a result,
many more repairs are performed at the central depot instead of
at the intermediate depot, which Fig. 14 shows. Since many
repairs are performed at the central depot, it is relatively
inexpensive to repair the subcomponents of those components
at the central depot too. Therefore, the discard costs in the
network with ten intermediate depots are lower than in the
network with two intermediate depots. In general, we see that
the more system locations there are, the more attractive it
becomes to buy resources and repair components instead of
discarding them. As a result, total costs increase not as much as
the number of system locations increases.
We conclude that, using our test instances, modelling the
repair network exactly brings cost reductions of almost 2% on
average for networks that are unbalanced in the locations. In
some cases, the cost reductions are over 7%, which means that it is
worthwhile to model the repair network exactly for unbalanced
networks. For networks that are balanced in the locations, cost
reductions are never higher than 1.2% in our test instances, which
means that it is doubtful whether the additional effort of
acquiring all inputs is worth it. In this case, aggregating all data
does not lead to much higher costs. However, we do note that if
the costs for spare parts are not equal in the whole network,
due to a difference in lead time between different parts of the
global repair network, it might be necessary to explicitly model
the repair network. However, more research is needed on the
integration of spare parts optimization into the LORA before this
question can be answered.
5. Conclusions and directions for further research
We have modelled the LORA problem as a minimum cost flow
model with side constraints. This formulation allows us to model
all kinds of extensions in an elegant manner. Such extensions
include repair probabilities, no-fault-found probabilities and
equipment with a finite capacity. Besides that, we have shown
that the LORA problem with all data aggregated per echelon level,
can be solved much faster using our formulation than it could
using existing formulations (a factor five on average for large
problem instances).
Our model allows us to explicitly model the repair network
In this paper we focussed on the LORA problem. The spare parts
stocking problem is usually solved after solving a LORA problem.
We noticed that this sequential approach can result in performing
many repairs at a central location and discarding many compo-
nents, which results in high spare parts costs. Therefore, it would
be very useful to solve these two problems simultaneously. This
might be done with an iterative approach using a LORA model and a
spare parts stocking model, but one integrated model could be
used as well.
Acknowledgments
The authors thank the anonymous referees for their comments,
which improved the quality of this paper. The authors gratefully
acknowledge the support of the Innovation-Oriented Research
Programme ‘Integrated Product Creation and Realization’ (IOP IPCR)
of the Netherlands Ministry of Economic Affairs and the support
of TRANSUMO.
Appendix A
In this appendix, we describe the random number generator
that we used in Section 4.2 to generate product structures, failure
rates and cost factors. We give a summary of all parameter
settings in Table 6.
In all tests, the system structure consists of 25 components at
the first indenture level (subsystems), 125 at the second level and
625 at the third level. Every second and third indenture
component is randomly attached to a lower indenture component
with equal probability per parent component. As a result, it may
occur that a first indenture component has zero subcomponents
(which also occurs in practice, e.g., at Thales Nederland). The
annual demand per subsystem is drawn from a uniform distribu-
tion on the interval [0.01; 1]. The annual demand for a higher
indenture component is a certain fraction of the demand for its
parent component. The fraction of demand for the parent
component that is due to a failure in the child component is
Table 6
Parameter settings.
Parameter Value(s) or range

instead of aggregating all information per echelon level, as is often # Components at indenture level 1 25
done in the literature. We have shown that with networks that are # Components at indenture level 2 125
unbalanced in the locations, substantial cost reductions can be # Components at indenture level 3 625
achieved by modelling the exact network (over 7% in our test
instances). If networks are balanced in the locations, then the Demand per subsystem [0.01; 1]
maximal costs reductions that can be achieved are much lower Fraction of demand for parent due to ½0:5=number of children;
(only 1.2% in our test instances), even if the network is child 1:25=number of children
Price of component [1,000; 100,000]
unbalanced in the costs. However, our research suggests that if
Repair costs (fraction net price) [0.1; 0.4]
lead times differ across the repair network, significant cost Discard costs (fraction gross price) [0.75; 1.25]
reductions may be achieved by modelling the exact network, Move costs (fraction gross price) 0.01
even for networks that are balanced in the locations. To be able to
analyse this, integration of spare parts optimization in the LORA is Resource costs [10,000; 1,000,000]
necessary. # Resources 10 & 25
We have shown that in some cases, only small cost differences Fraction of components requiring 0–1–2 0.7–0.2–0.1 & 0.25–0.5–0.25
exist between using the exact and the aggregated network, resources
Repair lead time at echelon level 1 1 month
although the decisions that are taken differ a lot. If there are
Repair lead time at echelon level 2 1 month
multiple ways to achieve almost the same total costs, there can be Repair lead time at echelon level 3 3 months
other, more qualitative reasons to choose for another solution Move lead time between two echelons 0.5 months
than the one with the lowest total costs. Future research could (1 & 2, 2 & 3)
lead to an approach that results in multiple alternative solutions Discard lead time 6 months
Safety factor 2
that differ a lot in terms of decisions, but lead to almost the same Holding costs 0.3
optimal solution value.
242 R.J.I. Basten et al. / Int. J. Production Economics 133 (2011) 233–242
drawn from a uniform distribution on the interval
½0:5=number of children; 1:25=number of children, with a max-
imum of 1. Summing the fractions of all children leads to a value
between 0.5 and 1.25. If this summation is lower than 1, then a
certain fraction of the failures in the parent can be repaired
directly, without replacement of a child component. If the
summation is higher than 1, then a certain fraction of the failures
in the parent leads to the replacement of more than one child
component.
For each component, we draw a net price, excluding the costs
of the children, from a shifted exponential distribution with shift
factor 1,000 and rate parameter l ¼ 7=ð100,0001,000Þ. In this
way, on average 1% of the components get a price higher than
h100,000. We draw a new price for these components to avoid
odd problem instances. The reason for our choice is that most
systems have a large diversity of items in price, but there are
considerably more cheap items than expensive items. The
cheapest items (in our case items with a price below h1,000Þ
are usually omitted from a regular LORA, because they are
discarded by default.
Using these prices, we calculate the variable costs as follows:
 repair costs as a fraction of the net component price are drawn
from a uniform distribution on the interval [0.1; 0.4];
 we recursively add the price of each child to its parent to get
the gross item price. We do this after calculating the repair
price, since repair of a parent means replacement of the child
that was defective and taking a decision for the child, thus
incurring costs for the child. Discarding or moving a parent
does not lead to a decision for the child components, and thus
not to incurring costs for the child components either;
 the discard costs as a fraction of the gross item price, including
children, are drawn from a uniform distribution on the interval
[0.75; 1.25]. The 100% would be just the costs of replacing a
defective component by a new one. However, there may be
additional disposal costs to satisfy legal requirements for
waste processing. On the other hand, some parts of a defective
component may be recycled or re-used. Therefore, we vary the
disposal costs around the price of a new component;
 the move costs are always 1% of the gross item price.
We perform tests with 10 and 25 resources. The price of each
resource is drawn from a shifted exponential distribution with
shift factor 10,000 and rate parameter l ¼ 7=ð1,000,00010,000Þ
in the same way as described above for the prices of the
components. As a result, prices vary between h10,000 and
h1,000,000. We randomly assign components to resources, such
that the percentage of components that needs 0, 1 and 2 resources
are 70%, 20% and 10%, respectively. To be short, we write ‘0.7–0.2–
0.1’. We also perform tests with settings ‘0.25–0.5–0.25’.
We also need to take spare part costs into account, see
Sections 1 and 4.2.1. As described in those sections, modelling
spare parts costs is not straightforward and should be item-based
since a system approach, such as the VARI-METRIC type approaches
(see, for example, Sherbrooke, 2004), is not possible. Our
approach is based on the difference in lead times that result from
different LORA decisions. We assume that repair at the system
location or at the intermediate depot takes a month, whereas
repair at the central depot takes three months. Discard (and
buying a new item) has a lead time of half a year. Moving a
component to a higher echelon leads to an additional lead time of
half a month. Using these lead times and a safety factor of 2, we
estimate spare parts inventory as being the safety factor times the
lead time demand. To compute the corresponding holding costs,
we use an inventory carrying charge of 30%. We tested other
approaches and safety factors, but found that this approach
generally leads to reasonable results.
References
Ahuja, R.K., Magnanti, T.L., Orlin, J.B., 1993. Network Flows. Theory, Algorithms,
and Applications. Prentice-Hall, Englewood Cliffs (NJ).
Alfredsson, P., 1997. Optimization of multi-echelon repairable item inventory
systems with simultaneous location of repair facilities. European Journal of
Operational Research 99, 584–595.
Barros, L.L., 1998. The optimization of repair decisions using life-cycle cost
parameters. IMA Journal of Mathematics Applied in Business & Industry 9,
403–413.
Barros, L.L., Riley, M., 2001. A combinatorial approach to level of repair analysis.
European Journal of Operational Research 129 (2), 242–251.
Basten, R.J.I., Schutten, J.M.J., Van der Heijden, M.C., 2009. An efficient model
formulation for level of repair analysis. Annals of Operations Research 172 (1),
119–142.
Basten, R.J.I., Van der Heijden, M.C., Schutten, J.M.J., 2010. Practical extensions to
the level of repair analysis. BETA working paper 304, submitted for publication.
Brick, E.S., Uchoa, E., 2009. A facility location and installation or resources model
for level of repair analysis. European Journal of Operational Research 192 (2),
479–486.
Deloitte, 2006. The service revolution in global manufacturing industries.
EDCAS, 2009. /http://www.tfdg.com/edcas.phpS, last checked on 9 May 2009.
Ferrin, B.G., Plank, R.E., 2002. Total cost of ownership models: an exploratory
study. Journal of Supply Chain Management 38 (3), 18–29.
Muckstadt, J.A., 2005. Analysis and Algorithms for Service Parts Supply Chains.
Springer, New York (NY).
Murthy, D.N.P., Solem, O., Roren, T., 2004. Product warranty logistics: issues and
challenges. European Journal of Operational Research 156, 110–126.
Oliva, R., Kallenberg, R., 2003. Managing the transition from products to services.
International Journal of Service Industry Management 14 (2), 160–172.
PRICE HL, 2007. /http://www.pricesystems.com/products/price_hl.aspS, last
checked on 25 July 2007.
Saranga, H., Dinesh Kumar, U., 2006. Optimization of aircraft maintenance/support
infrastructure using genetic algorithms—level of repair analysis. Annals of
Operations Research 143 (1), 91–106.
Sherbrooke, C.C., 2004. Optimal Inventory Modelling of Systems. Multi-echelon
Techniques, second ed. Kluwer, Dordrecht, The Netherlands.