Académique Documents
Professionnel Documents
Culture Documents
To cite this article: Marco Tantardini, Alberto Portioli-Staudacher & Marco Macchi (2014) A model for considering the impact
of rescheduling planned maintenance activities in a maintenance service contract, Production Planning & Control: The
Management of Operations, 25:3, 241-259, DOI: 10.1080/09537287.2012.665094
Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
Production Planning & Control, 2014
Vol. 25, No. 3, 241259, http://dx.doi.org/10.1080/09537287.2012.665094
Maintenance services are becoming increasingly relevant and both service providers and users are paying more
Downloaded by ["Queen's University Libraries, Kingston"] at 05:15 06 January 2015
attention to cost and service level. Nonetheless, until now very little attention has been devoted for understanding
and evaluating rescheduling costs which can be quite relevant, particularly if maintenance is outsourced to
external providers. This article proposes a model to measure the costs for rescheduling maintenance
interventions, thus allowing to take them into consideration both when defining a service contract and when
running it. The model has been developed after a wide empirical analysis deployed in two phases: a first
exploratory phase aimed at identifying the variables of the rescheduling problem by interviewing the main
maintenance service providers in Italy; the second, a confirmatory phase has been developed validating the model
using data from real industrial setting and analysing the results with the managers who supplied the data.
Keywords: outsourcing; maintenance service contracts; rescheduling; cost model
models, used to solve an optimisation problem con- (Taracki et al. 2006), and the strategic importance of
sidering the degradation of production equipment and coordinated actions between the producer and the
simultaneously minimising the costs incurred in the service provider (Kumar et al. 2004), are the main
production and maintenance domain, such as produc- motivations for focusing our research on improving
tion, setup, inventory, labour, maintenance and quality the interaction and the relationship between these two
costs (see, e.g. Weinstein and Chung 1999, Pistikopolos players and for selecting this problem scope.
et al. 2001, Aghezzaf et al. 2007, Aghezzaf and Najid Furthermore, this research originates from the request
2008); (ii) planning models, where the stocking of of an important provider of maintenance services to
production material is included as a buffer to absorb better investigate how the rescheduling problem is
the demand variation and the changes in the produc- tackled in real companies: this is another proof of
tion rates subsequent to the stoppages of equipment interest from the industrial field.
for maintenance activities (see, e.g. Chelbi and In the remainder of this article, outsourcing in
Ait-Kadi 2004, Chelbi and Rezg 2006, Chelbi et al. maintenance is first introduced with its main features;
2008, Regz et al. 2008, Radhoui et al. 2009); (iii) sched-
Downloaded by ["Queen's University Libraries, Kingston"] at 05:15 06 January 2015
the cost model is then put into this context and briefly
uling models, aimed at optimising typical operational presented in its intended use (Section 2). The
performances, like the flow time, tardiness, makespan rescheduling problem is analysed in more detail
and availability, subject to the constraints created by (Section 3), assuming the perspective of the provider
the stoppages scheduled for maintenance reasons of maintenance services: the results of an exploratory
(Lee 1999, Qi et al. 1999, Lee and Chen 2000, case study analysis are herein shown to identify the
Kubiak et al. 2002, Cassandry and Kutanoglu 2003, variables of the problem. As a follow up, the cost
2005, Aggoune 2004, Crespo Marquez et al. 2006, model is formulated and discussed in its use in
Chen 2007, 2008; Kuo and Chang 2007, Ruiz et al. practice (Section 4). The model is then validated
2007; Liao and Tsai 2009). Single machine and flow (Section 5) based on the data sets taken from different
shop are the most common systems studied by the industrial settings. Concluding remarks (Section 6)
literature; moreover, it is worth observing that only eventually discuss the implications that can be
specific models, such as Yulan et al. (2008), consider elicited, from our work, for industrial practice and
cost as part of a multiple objective function of the joint scientific research.
scheduling problem.
Notwithstanding this wide literature on joint deci-
sions between production and maintenance, to the best
of our knowledge, no one has expressly discussed in the 2. Problem scope
past the problem of rescheduling planned maintenance The recent trends in industry show that outsourcing in
interventions and of the resulting inefficiencies. To this maintenance has become a common practice in
end, we propose a cost model for assessing the different sectors (Taracki et al. 2006, Garg and
rescheduling inefficiencies. The model is developed Deshmukh 2010).
starting from the research work on MRP nervousness According to Harland et al. (2003), outsourcing is
presented, for example, in Ho et al. (1995), Ho and motivated by the objective to involve specialists to
Carter (1996) and Kadipasaoglu and Sridharan (1997). provide competences, technologies and resources.
These papers tackle the topic of rescheduling produc- Basically, three fundamental factors bring companies
tion orders generated by MRP, and thus represent a to use outsourcing in maintenance: (i) cost reduction,
good basis for developing our research into the issue of since specialised providers are more experienced, have
rescheduling maintenance orders. The final outcome of more sophisticated equipments and decision models,
this article is a model that is an analytical approxima- hence they can provide better service efficiency;
tion of the rescheduling costs. The model can be used (ii) more skills, since using external resources is a
to support cost-based decisions within the maintenance way to get rapidly new skills via qualified people and
scheduling system. For developing the model we have specific instruments; (iii) a higher service level, since
investigated, through case studies, the operational this can be linked to a system of rewards/penalties
variables typical of the scheduling system and their established in the contract in a so-called bonus/malus
effects on costs. The cost model is proposed as a formula and measured on Service Level Agreements
supporting tool for rescheduling decisions jointly (SLAs) for which the provider is duly responsible.
achieved, in an inter-organisational context, by the The development of appropriate contracts is essen-
third party, providing the maintenance service, and the tial to take advantage of these factors. Indeed,
customer. The recent trends observed on outsourcing according to Panesar and Markeset (2008), a good
in maintenance as a common industrial practice contract is needed in order to create an environment
Production Planning & Control 3
243
Operative/operational maintenance Includes all the activities at operational level, such as equipment repair, intervention
preparation, intervention reporting and so forth.
Maintenance management Includes all the activities at management level, such as work order request management,
work order assignment, data storage, resources management, and so forth.
Contractual management Includes all the activities needed to transfer administrative information, contracts and
documents between customer and provider.
Data management Includes all the activities necessary for the acquisition and transfer of equipment data
from the customers site to the providers service workshop, for their further analysis
in maintenance engineering, condition monitoring, and so forth.
Spare parts management Includes the acquisition and storage of spare parts, as well as analysis on data regarding
spare parts logistics.
Maintenance engineering Includes all the activities aimed at improving performance of the maintenance contract
Downloaded by ["Queen's University Libraries, Kingston"] at 05:15 06 January 2015
by leveraging on preventive (cyclic time based and condition based) maintenance and
proactive maintenance.
KPI monitoring Includes the monitoring at different levels of analysis (KPIs at operative, tactical, and
strategic levels) of maintenance activities.
Training activities Includes the activities offered to the customers maintenance personnel in order to
enable the transfer of information and knowhow required by maintenance
interventions.
facilitating the execution of the maintenance activities A proper monitoring of key performance indicators
planned in the outsourcing scope. A general prerequi- (KPIs) is a strategic process for the contract success.
site for the success of outsourcing is the proper It is needed, first of all, in order to control how far the
establishment of contractual agreements, as they are contract is meeting the established SLAs and, as a
the instruments which will allow both customer and consequence, to calculate related penalties/rewards.
service provider to share what has to be done in the Besides, measuring maintenance activities allows for
outsourcing scope (Kumar et al. 2004). In particular, checking whether management practices are properly
contract clauses have to be established in order to deployed in the processes and work flows are linked to
define the equipment for the maintenance services to be the contract; a positive measurement is a leading
provided; services are then declared (services can range indicator of the quality of the contract execution and,
in a very wide spectrum from, e.g. field services to do as a consequence, an outlook on its future, positive
repairs, expert assistance for diagnostics; logistics results.
support; advanced training; etc.); finally, the job It is also strategic, especially in performance-based
responsibilities for each party, maintenance processes contracting (Kim et al. 2007), that decisions made in
and related work flows are agreed upon with regard to the maintenance management process are coordinated
the equipment and services established in the contract between the customer of the service, and the provider
(Kumar et al. 2004). of the service: indeed, maintenance work order assign-
Table 1 presents, with no claim of exhaustiveness, a ments to proper resources and times should result from
range of candidate processes for maintenance out- a joint decision, in order to take into account the
sourcing, and it is the synthesis of different parts respective exigencies issued by the customer (i.e.
adapted from different authors at the scientific and production exigencies) and the provider (i.e. mainte-
international levels. A main inspiration can be the nance exigencies).
above-mentioned work by Kumar et al. (2004). Also, Considering these two relevant processes, we pro-
we think that it is important to refer to technical pose a cost model in order to measure the inefficiencies
publications, because they provide contributions con- resulting from the modifications of a previously
cerning the practices issued in industry for contract committed maintenance schedule. The model proposed
setup and execution; in this regard, we can mention for allows, on the one hand, for a more effective costs
example Gomez et al. (2009). Other references can be control on the part of the maintenance service
sourced from the literature on after sales management provider; on the other hand, it enables to quantify
(see, e.g. Legnani et al. 2009), discussing the configu- more precisely possible economic compensation for the
ration of assistance processes according to the after flexibility exhibited in such rescheduling situations by
sales strategy of a company. the maintenance service provider.
4
244 M. Tantardini et al.
viewpoints on the rescheduling problem. This is the F After sales department of a metalwork
reason why we have included pure service companies multinational mass producer
providing maintenance services, manufacturing firms G After sales department of a company produc-
ing ETO equipment for processing
providing maintenance as after sales services and a industries
consulting company in outsourcing maintenance inter-
ventions in our case study analysis (this company
advises manufacturing companies on how to structure
the outsourced maintenance contracts). human resources are involved in carrying out mainte-
We have decided to consider all these groups of nance interventions for different customers in different
companies because all of them could be affected by the plants, rescheduling an intervention may result in the
problem of rescheduling planned maintenance inter- need to modify other interventions. Consequently,
ventions and could be interested in understanding how rescheduling is more expensive as the maintenance
to better manage such an issue. In fact, all these groups intervention has to be rescheduled in periods showing
of companies are interested on the one hand in defining already committed interventions for the maintenance
the fair pricing for the maintenance service contract resources provided to operate a given contract, and
and on the other hand are interested in being able to when such interventions require highly skilled techni-
evaluate the inefficiencies brought by the modification cians who are often more saturated because of their
of already committed interventions. Moreover, includ- higher costs. Should it not be possible to obtain a
ing all these groups of companies could make easier for skilled resource in-house for the period required, it is
us in assessing the problem of rescheduling in a broad then necessary to hire competences from third parties,
sample of companies and business context, with the which are very expensive. Besides, inefficiencies deriv-
ultimate goal to highlight all the cost components that ing from rescheduling also affect the management of
are worth to be considered in the rescheduling cost technical materials (spare parts, tools and equipment).
model. For example, when an anticipation for an intervention
Among these three groups of companies, we have is required, the materials might not be available in
decided to study only multinational companies with
stock, and this could cause a costly expediting. On the
hundreds of employees because we believe that within
contrary, when an intervention has to be postponed,
these companies it is more likely to highlight possible
the service provider may require extending the hiring of
best practices and thus an overview of the industrial
some specific equipment necessary for the maintenance
state-of-the-art on the topic of rescheduling mainte-
intervention.
nance interventions. Specific case studies were then
Last but not the least, the manager underlines that
selected from a list of companies based in the North of
it is important to know about the rescheduling request
Italy that we know they put emphasis and invest in the
with an adequate anticipation: a low time to the
development of best practices in maintenance. Table 2
intervention that is going to be rescheduled causes high
presents the sample of companies interviewed.
inefficiencies, especially in the case of anticipation.
Even if rescheduling is perceived as a critical issue,
the associated costs are currently neither measured
3.1. Company A (apart from the costs of hiring external resources) nor
The manager of company A reports on the high expressly considered in the maintenance contract.
relevance of the rescheduling problem. Since the Thus, the service provider currently has no levers to
Production Planning & Control 2455
demonstrate such inefficiencies to the customer and to spare parts management is adversely affected by
obtain an acknowledgement for its flexibility whenever rescheduling.
a change to the schedule is requested. Nonetheless, to This leads to a list of inefficiencies that are usually
reduce the inefficiencies, the manager of company A under-estimated or even not taken into account in the
sometimes agrees with the customer the possibility for real practice of the company. Indeed, the manager
the providers company workforce to carry out addi- states that the rescheduling costs are not communi-
tional activities (e.g. set-up) in order to avoid a possible cated to the customer and are not currently inserted
under-saturation, induced by a rescheduling decision. into the contracts because they are difficult to be
measured. However, to reduce the inefficiencies, the
manager states that sometimes he defines in the
3.2. Company B contract an upper bound for rescheduling decisions:
the customer has to pay a lump sum roughly defined
Also the manager of company B reports on the
criticality of rescheduling, usually occurring as a on the spot situation for each additional change over
a defined limit.
Downloaded by ["Queen's University Libraries, Kingston"] at 05:15 06 January 2015
and availability of required materials or equipment. In highly saturated, the resulting rescheduling costs may
particular, the inefficiencies resulting from reschedul- increase dramatically.
ing become more critical as the intervention to be In fact, in the above situations, rescheduling one
rescheduled is close to the current time period. More activity might be very complex due to the constraints
precisely, the manager of company E states that the from the other interventions committed and might
more you need to anticipate the intervention, or the likely require rescheduling other activities in order to
closer the intervention to be rescheduled is to the cur- obtain a feasible plan, or might also require hiring
rent date, the more likely the reschedule required could temporary workers.
be unfeasible or anti-economical. The manager of company F does not measure the
The manager of company E affirms that the rescheduling costs precisely. He only uses a rough
flexibility to respond to the production exigencies is estimation for remarking to the customer that the high
appreciated by the customer, although not formally flexibility for rescheduling has a cost for the provider.
acknowledged. On the other hand, company E neither This is a lever in order to obtain an economic
compensation when the number of required resched-
Downloaded by ["Queen's University Libraries, Kingston"] at 05:15 06 January 2015
the customer: company G only requires an economic waste transportation tickets already
compensation for the direct costs of manpower when bought.
the maintenance activity is shifted to the week-end. In (IV) Anticipation of the rescheduling notice.
Table 3 we have summarised the main findings in the When the need to reschedule an inter-
case study analysis. vention is communicated with poor
advance, then higher costs for both
workforce and materials are to be
3.8. Summary of the case study analysis expected because it is required to
The multiple case study analysis highlighted several re-arrange the intervention in a short
interesting findings, enlisted in the remainder. time with a limited number of levers to
manage efficiently such rescheduling.
(1) Rescheduling is a very critical problem in all (V) Number of shifting periods. The higher
the companies that we have interviewed: it is the number of shifting periods, the higher
frequent and the associated costs are not the cost for both workforce and spare
Downloaded by ["Queen's University Libraries, Kingston"] at 05:15 06 January 2015
8
Table 3. Summary of case studies.
248
Are rescheduling Very much Very much Very relevant and Very relevant and Relevant and Generally yes, but Yes, about 25% of
problems relevant? particularly frequent frequent depend on the the planned
frequent intervention to interventions are
be rescheduled rescheduled
Which costs are Equipments and Equipments and Equipments and Manpower Equipments and Manpower Equipments and
affected? materials materials materials materials materials
Manpower Expected costs of Manpower Manpower Manpower
corrective main- Administrative Administrative
tenance staff activities
Manpower and
manpower-
related
(e.g. travels)
Administrative
staff
Which are the main Presence of other Presence of other Number of shift Presence of other Time to the Number of shift Number of shift
rescheduling opera- interventions interventions periods interventions intervention periods periods
tional variables? already commit- already commit- Time between already commit- Number of shift Number of inter- Materials
ted in the ted in the two consecutive ted in the periods ventions to be required
period period interventions period rescheduled Time to the
Highly skilled Number of shift Periods with Time between intervention
technicians periods higher costs for two consecutive
required Time to the carrying out interventions
Number of shift intervention interven- Typology and
periods tions (e.g. complexity of the
Time to the weekends) interventions to
intervention External be rescheduled
resources External
M. Tantardini et al.
required resources
required
Are costs measured No (apart from the No No, because they No No Not measured pre- Costs are measured
and enclosed in the costs of hiring are difficult to cisely, not but not enclosed
contract? external be measured included into the in the contract
resources) contract apart the direct
costs of mainte-
nance during
week-ends
How the provider Provider sometimes No specific actions Sometimes an When the number Manager requires
manage as to asks to carry out upper bound for When the resche- Sometimes it is of rescheduling that the notice
reduce the resche- additional activ- rescheduling duling date is agreed the mini- activities for a reschedul-
duling ities (e.g. set-up) decisions is set: unfeasible for the mum anticipation required ing decision
inefficiencies? in order to avoid for further provider, a new to issue a request increases too arrives at least
a possible reschedules date is agreed for rescheduling much the man- two weeks in
under-saturation required the upon maintenance ager uses a advance to the
induced by a customer has to Manager sched- activities rough estima- scheduled event
rescheduling pay a fixed sum ules activities for Provider some- tion of resche-
decision 80% of the time times asks to duling costs for
of human carry out addi- remarking the
resources avail- tional activities cost of the
able, the rest is (e.g. set-up) in flexibility
kept free to order to avoid
absorb the under-saturations
variations
Production Planning & Control 2499
In our model we have only considered the real costs function with the number of interventions to be
that are related to the resources provided in the rescheduled.
maintenance service contract and that are therefore
both easier to be expressed concretely and easier to be
allocated to the responsibilities of the involved parties. 4.2. Maintenance materials
With such logic, the failure risk cost associated to the Rescheduling also affects the technical materials costs.
shift of an already committed event is, for example, not This occurs whenever the maintenance activity requires
considered in our model for various reasons. First, a replacement of components (spare parts) or the use
such a cost could be difficult to be measured and be of special equipment/tools (e.g. cranes, scaffoldings).
shared by the parties, because it is based on a We assume these costs to be directly proportional
probabilistic approach (see, e.g. Fumagalli et al. to the number of periods of anticipation (or postpone-
2010). Second, a failure risk cost becomes a real cost ment): if the temporal shift gets higher, the costs
only as the failure happens and when the failure increase (cf. Section 3.1 operational variable V). This is
Downloaded by ["Queen's University Libraries, Kingston"] at 05:15 06 January 2015
happens after a rescheduling request, it could be not expected to represent well the following cases: (i) the
that easy to attribute the proper responsibilities. costs for urgent deliveries due to expediting, which can
Moreover, for the sake of simplicity, we have be considered proportional to the anticipation; (ii) the
considered all the various maintenance interventions as warehousing costs for equipment/spare parts or the
having the same degree of criticality/complexity. costs of extension for the hiring of special equipment/
Consequently, despite the suggestions from the tools, proportional to the postponement.
manager of company F, rescheduling any intervention Costs are also dependent on the typology of
has been considered in our model to have the same material. Finally, they are set to be inversely propor-
unitary cost. Therefore, the operational variable II tional to the time to the intervention (cf. Section 3.1
(Section 3.1) was not included in our proposal. operational variable IV): if the rescheduling is issued
We have also considered, for the sake of simplicity, with a certain anticipation, managers may have more
that performing maintenance activities has the same levers in order to reduce the inefficiencies due to
cost in each planned period. Therefore, despite the rescheduling.
suggestions from the manager of company D, we have
not identified more expensive or cheaper periods in
which to reschedule maintenance interventions. The
4.3. Maintenance workforce
operational variable I (Section 3.1) was thus not
included in the proposed model. Maintenance workforce costs are associated to the
inefficiencies due to the rescheduling of human
resources. We assume that such costs are assessed
mainly when the human resources of the providers
4.1. Administrative staff company are shared among different customers.
Every time a rescheduling is requested, the planner of In fact, in this case, the rescheduling decision in
the service provider needs to re-process maintenance order to respond to a request of one customer affects
requirements and resource availability, to define a new the provision of such common resources to provide the
feasible schedule. service to another customer, thus finally impacting on
The unit cost value depends on the fact that the service duties of different contracts.
resources are shared with different customers/plants The unitary costs increase as the workforce is more
(cf. Section 3.1, operational variable VI). The more the specialised and skilled. Besides, we make a distinction
resources are shared with different customers, the whether it is required to anticipate or to postpone the
higher the cost because of the constraints increase. intervention. When it is required to anticipate an
Moreover, such administrative cost associated with the intervention, we suppose the costs to be directly
rescheduling activity is mainly related to the number of proportional to the number of required periods of
rescheduling actions (operational variable III in anticipation (cf. Section 3.1 operational variable V),
Section 3.1). because it is likely that the resources are already
We assume that the same unit cost may apply to committed to other interventions. Instead, when an
both anticipating and postponing events. Nonetheless, intervention has to be postponed, we have considered
it is worth considering the complexity of rescheduling that the resulting cost is a fixed value, not related with
multiple interventions, in order to finally model a the number of shifting periods. Thus, the number of
more expensive cost if compared to rescheduling shifting periods is not regarded as influent on
a single intervention. Thus, we set a linear cost the cost formulation because, with postponement,
10
250 M. Tantardini et al.
Anticipation Postponement
Notes: Here:
j 1, . . . , J
workstations (subject of the maintenance service contract)
" 1, . . . , E
maintenance interventions
arac fixed unitary cost for administrative rescheduling
Na, Np number of anticipated (Na) or postponed (Np) interventions
acm"j, pcm"j materials costs, related to anticipating (postponing) one period the "-th planned maintenance intervention on
Downloaded by ["Queen's University Libraries, Kingston"] at 05:15 06 January 2015
workstation j.
am"j, pm"j smoothing coefficient for materials costs related to anticipating/postponing the "-th planned maintenance
intervention on workstation j.
A"j, P"j number of anticipation/postponement periods for the "-th planned maintenance intervention on workstation j.
acw"j, pcw"j workforce costs related to anticipating (postponing) one period the "-th planned maintenance intervention on
workstation j.
aw"j, pw"j smoothing coefficient for workforce costs related to anticipating/postponing the "-th planned maintenance
intervention on workstation j.
P"j binary variable: it is equal to 1 in case the "-th maintenance intervention on workstation j is postponed,
otherwise 0.
the workforce is likely to be idle or featuring a lower pw"j defined in Equation (4) allows for controlling the
workload. In fact, when postponing an intervention smoothing effect of the rescheduling costs. The value
conflicts with other activities already scheduled rarely of such coefficients depends on how much the com-
emerge. Consequently, it is reasonable to measure munication to reschedule the intervention is noticed in
solely a fixed cost resulting from the rescheduling advance to the planned intervention and on the two
inefficiencies caused by the variation of the human parameters ( and D) that enable to adapt the cost
resources program (e.g. the resource has to modify its model to each particular provider.
travelling plan to the sites). tm"j T1
Similarly to materials, the anticipation/postpone- D 1 "j
"j "j , tm"j , T, D 1
4
ment costs are also considered to be inversely propor-
tional to the time to the intervention (cf. Section 3.1, where:
operational variable IV): the longer the time to tm"j period when the "-th maintenance interven-
intervention, the lower the impact of its modification. tion has been planned in the previous
schedule
T period when the rescheduling decision is
4.4. Mathematical formulation issued
"j parameter, ranging from 0 to 1, measuring
Table 4 reports the mathematical formulation for the
the providers flexibility in either anticipat-
different cost components. An important element of
ing or postponing interventions
the formulation of costs for maintenance material and
D parameter, ranging from 1 to tm"j (T 1),
workforce is the presence of some coefficients for
defining the number of periods during
smoothing, introduced in order to consider the effect of
which rescheduling costs shall be assumed
issuing a request to reschedule with proper
as steady
anticipation.
We have exploited the research work on MRP As "j tends to 1 (i.e. high flexibility), the smoothing
nervousness presented in Ho et al. (1995), Ho and coefficient in Equation (4) tends to 0. In such a
Carter (1996) and Kadipasaoglu and Sridharan (1997), situation it is possible to avoid extra-costs for the
in order to define the values of the smoothing service provider, thus it is not costly to reschedule the
coefficients am"j, pm"j, aw"j, pw"j. The formula maintenance interventions. Besides, as the parameter D
for the smoothing coefficients am"j, pm"j, aw"j, decreases, the smoothing coefficient is rapidly cut.
Production Planning & Control 11
251
Table 5. Summary of information required to run the increases as time elapses from the last maintenance
model. intervention this enables to consider the growing risk
of failure. For modelling failure rates we used the same
Standard Reschedule
costs of parameters of logic as in Pistikopoulos et al. (2001), that is to have
the service the customer constant and increasing rates in different periods.
contract request More precisely, maintenance costs express the total
costs of the maintenance service, divided among:
Administrative staff arac Na
(Equation (1)) arac Np
(1) the cyclic preventive maintenance during the
Maintenance materials acm"j am"j A"j weekly timetable, (2) the cyclic preventive maintenance
(Equation (2)) pcm"j pm"j P"j during the week-end and (3) the corrective mainte-
Maintenance workforce acw"j aw"j A"j nance, based on a stochastic failure rate. The model
(Equation (3)) pcw"j pw"j P"j also considers the following production costs: (1) setup
costs, (2) manufacturing costs, (3) inventory costs,
Downloaded by ["Queen's University Libraries, Kingston"] at 05:15 06 January 2015
Average cost for a preventive 11,000E . impossibility to carry out the maintenance
maintenance intervention in regular
time intervention in the planned period
Average cost for a preventive 13,000E (e.g. because reworks are required, because
maintenance intervention in over time tests are required on new products, etc.);
Average cost for a corrective 17,000E . increase of demand in periods coinciding with
maintenance intervention
already scheduled intervention.
Average time for a preventive 16 h
maintenance intervention Rescheduling activities result in high inefficiencies
Average time for a corrective 72 h
maintenance intervention
in validation company 1 for three operational reasons:
. maintenance planners are highly saturated
and paid high salaries, so their time should
be used in an efficient way;
Production can be grouped into two families . technicians come from far away; the provider
(X and Y), quite similar one to the other apart from of maintenance services is in fact geographi-
the raw materials used. This is the reason why family Y cally far from validation company 1;
is 10% more expensive to produce than family X. . the preventive maintenance interventions are
Each family is produced on alternate weeks with quite complex, and they require a highly
batches of 2.5 million pieces. Set-up time is 4 h and 5 skilled workforce.
operators are required to change from family X to Therefore, the fixed costs due to administrative
family Y. Set-up cost is 1000E/set-up (5 operators * 4 h activities required for rescheduling are quite high
* 50E/h). In Table 6 we present the production data, (i.e. 400E regardless of the number of periods of shift
while in Table 7 we present the maintenance data. required), while the costs related to material and
Maintenance activities are outsourced to the produc- human resources may vary significantly in the case of
tion line manufacturer. anticipations or postponements.
A preventive maintenance intervention takes 16 h Anticipating interventions creates higher inefficien-
(12 h of maintenance plus 4 h for restarting produc- cies than postponing interventions. Nonetheless, such
tion). A corrective maintenance intervention instead inefficiencies also depend on the number of shifting
takes 72 h because providers technicians are required periods. In Table 8 we summarise the rescheduling
to reach the customer site. costs in the examined case.
With the data provided by the manager inter- It is then to be noticed that rescheduling costs
viewed, we ran the joint model for production and further depend on the time from the notice of the
maintenance on a planning horizon of 28 weeks in rescheduling and the maintenance planned period. In
order to simulate at least 24 weeks (i.e. 6 months) at a this situation it is fundamental to obtain the notice as
steady state. soon as possible. Therefore, when setting the param-
The joint master schedule produced three interven- eters for the cost smoothing coefficients (in particular,
tions: 1 in week 8, 1 in week 15, 1 in week 22, all of for and D), the manager took into account all these
them required during regular working time. From such aspects. D was set equal to 4 with a week as the unit
a solution, we have simulated different possible period in order to maintain rescheduling costs steady in
scenarios that require rescheduling activities. The the month. A graphical analysis of the different values
14
254 M. Tantardini et al.
Event Costs
1 In week 6 the impossibility to carry out the maintenance event in week 8 arises. MC: 57,310E
Old schedule: w8w15w22 RC: 7286E
New schedule: w7w14w21 Impact:12.7%
2 In week 6 the planned event at week 15 must be anticipated. MC: 57,310E
Old schedule: w8w15w22 RC: 1880E
New schedule: w8w14w22 Impact: 3.3%
3 In week 13 the impossibility to carry out the maintenance event in week 15 arises. MC: 39,510E
Old schedule: w15w22 RC: 6280E
New schedule: w14w21 Impact: 15.9%
4 In week 6 an increase of 10% in demand volume for period 8 comes up. MC: 57,310E
Old schedule: w8w15w22 RC: 1880E
New schedule: w8w14w22 Impact: 3.3%
Downloaded by ["Queen's University Libraries, Kingston"] at 05:15 06 January 2015
5 In week 6 an increase of 20% in demand volume for period 8 is recorded. MC: 57,310E
Old schedule: w8w15w22 RC: 7286E
New schedule: w7w14w21 Impact: 12.7%
6 In week 6 an increase of 30% in demand volume for period 8 is registered. MC: 58,160E
Old schedule: w8w15w22 RC: 12,766E
New schedule: w6w13w21 Impact: 21.9%
7 In week 6 an increase of 10% in demand volume for periods 7 and 8 arises. MC: 57,310E
Old schedule: w8w15w22 RC: 2907E
New schedule: w9w16w23 Impact: 5.1%
8 In week 6 an increase of 20% in demand volume for periods 7 and 8 is registered. MC: 58,160E
Old schedule: w8w15w22 RC: 10,280E
New schedule: w6w14w22 Impact: 17.67%
helped the manager to set the proper value of . integration of some mechanism in the maintenance
In particular: management process; this may help improving the
coordinated decisions between Customer (production)
. anticipation material 0.3;
and Provider (maintenance) even at the rescheduling
. anticipation human resources 0.5;
moment. Referring specifically to the data of this
. postponing material 0.6;
particular case, for example, assume that the two
. postponing human resources 0.8.
parties have agreed an upper limit in a default decision
In Table 9 we present the results from the analysis in the contract for rescheduling at 10% of impact on
of the different scenarios tested for the case, where in the total maintenance costs: then, only scenarios 2, 4
the column costs we report maintenance costs (MC) and 7 may be acceptable by default. This means that
and rescheduling costs (RC). MC represent the sum of other requests might be rejected (or subject to further
the preventive maintenance costs and the expected negotiations before deciding on the reschedule), with
corrective maintenance costs from the joint planning the end effect of avoiding too frequent changes and,
model utilised for generating the schedule. Instead, RC finally, of keeping a better stability in the maintenance
represent the expected rescheduling costs to change program. Furthermore, highlighting a situation with
from the old schedule to the new schedule, calculated very high rescheduling costs may also suggest to the
through our cost model. Therefore, RC are not two parties to explore different and likely cheaper
enclosed in MC. alternatives, for example by postponing a production
The impact of rescheduling costs in validation order or by redesigning maintenance activities distri-
company 1 is critical since it varies a lot (from 3.3% to bution over the following interventions.
about 22%). Not taking into account such costs in this
case actually means not controlling a relevant part of
the expenditures. Thus, it seems to be important for 5.2. Validation company 2
this company to consider such costs in the maintenance This company is a medium-sized manufacturing firm
service contract. that employs about 120 people. The annual turnover is
In particular, having the knowledge of the costs due around 15 million euros and the core business is to
to rescheduling inefficiencies may help envisioning the produce and sell valves for tires (for both cars and
Production Planning & Control 15
255
Table 10. Vulcanisation department production data. Table 11. Vulcanisation department maintenance data.
trucks). Moreover, the company produces manome- Table 12. Rescheduling costs vulcanisation department.
Downloaded by ["Queen's University Libraries, Kingston"] at 05:15 06 January 2015
Scenario Costs
Hence, the acceptance of a rescheduling request flexibility to follow production for all the changes
issued by the customer may be assessed with these required (see, e.g. in Section 3, our companies B and G
numbers: referring to the data of the case, assume, for of the exploratory case study analysis).
example, that Customer and Provider have agreed in
the contract on an upper limit of 5% impact, as
threshold to accept by default the rescheduling request;
6. Conclusions
only scenarios 1, 4 and 5 may be executed under this
agreement. Rescheduling maintenance interventions is a frequent
As a conclusion to this case, then, we can confirm issue in maintenance service contracts. Nonetheless,
the results of validation company 1. Indeed, consider- today the providers of maintenance services hardly
ing expressly rescheduling costs is important also in measure the inefficiencies deriving from rescheduling in
validation company 2 for supporting the decision- terms of costs.
making process about rescheduling events. The higher In the current mindset of managers, production has
the generated inefficiency, the higher the need for the highest priority and maintenance providers have to
highlighting a rescheduling alternative. be flexible enough to modify the schedules as produc-
Moreover, it is worth observing that validation tion requires. Provider flexibility is thus very appreci-
company 2 is different from validation company 1. ated by the customer (production), but there is no
At the beginning of the study in validation company 2, robust scheme to correspondingly acknowledge an
the operational reasons for being efficient seemed quite economic compensation.
important but were dominated by the need to provide This happens even if rescheduling costs are esti-
quick repairs in order to reduce the impact of mated to be an important quota of the maintenance
production losses during corrective maintenance. costs. For example, in the validation case studies
At the end of our study, we have shown that the presented in this article rescheduling costs can range
inefficiencies from rescheduling are significant indeed from a 2.3% to 22% of the total maintenance costs of
and thus it is very important not to neglect them. This the master schedule. This underlines the importance of
conclusion can be used as a general remark for all considering them. Hence, some managerial implica-
those companies recognising that the problem of tions have to be elicited from our research.
maintenance efficiency exists but declaring, at the First of all, the service provider should consider
same time, that maintenance has to provide all the such rescheduling costs since they represent a relevant
Production Planning & Control 17
257
part of its expenditures. Not measuring well these costs flexibility of maintenance service are achieved, we
means neglecting an important quota of the total costs envision that this theory may be considered in the field
of the maintenance service contract and possibly of life cycle costing (LCC) of physical assets. Using
wrong understanding of the profitability of the LCC, we expect to discuss the benefits (reduction of
contract. operating costs) of investing in an improved flexibility
Therefore, enclosing rescheduling costs in the of the maintenance services.
maintenance contracts, in the negotiations between Concluding, in this article we have demonstrated
providers and customers and, last but not least, in the that neglecting rescheduling costs is a bad, but very
KPI monitoring process, could allow for the possibility frequent practice for a maintenance service provider.
to provide a new support for decision making and the For future works, we expect new studies investigating
information needed to calculate the economic com- the problem of rescheduling planned maintenance
pensation for provider flexibility. As the formulation interventions, both by enlarging empirical proofs and
of the cost model is clear and has been agreed upon by investigating the adoption of the model in the theory of
the contracting parties, this can be the base ground for LCC of physical assets.
Downloaded by ["Queen's University Libraries, Kingston"] at 05:15 06 January 2015
Executive MBA at MIP the Business School of Politecnico Chelbi, A. and Rezg, N., 2006. Analysis of a production/
di Milano where he is responsible of the operations and inventory system with randomly failing production unit
quality management area. He leads research and consultancy subjected to a minimum required availability level.
projects on operational excellence in manufacturing and International Journal of Production Economics, 99 (12),
service companies in Italy and Europe. 131143.
Chen, W.-J., 2007. Minimizing total flow time and maximum
Marco Macchi is a Researcher at tardiness with periodic maintenance. Journal of Quality in
Politecnico di Milano, currently teach-
Maintenance Engineering, 13 (3), 293303.
ing modelling of production systems
and logistics and design and manage- Chen, J.S., 2008. Scheduling of non-resumable jobs and
ment of production plants, Vice flexible maintenance activities on a single machine to
Director of the Executive Master on minimize makespan. European Journal of Operational
Industrial Maintenance Management, Research, 190 (1), 90102.
scientifically responsible of the Crespo Marquez, A., Gupta, J.N.D., and Ignizio, J.P., 2006.
Observatory on Technologies and Improving preventive maintenance scheduling in semicon-
Services for Maintenance, Secretary of the IFAC Working ductor fabrication facilities. Production Planning &
Downloaded by ["Queen's University Libraries, Kingston"] at 05:15 06 January 2015
Kuo, Y. and Chang, Z., 2007. Integrated production processes of the industrial enterprises, MITIP 2008,
scheduling and preventive maintenance planning for Prague, 148154.
a single machine under a cumulative damage failure Qi, X., Chen, T., and Tu, F., 1999. Scheduling the
process. Naval Research Logistics, 54 (6), 589708. maintenance on a single machine. Journal of the
Lee, C.Y., 1999. Two-machine flow shop scheduling with Operational Research Society, 50 (10), 10711078.
availability constraints. European Journal of Operational Radhoui, M., Rezg, N., and Chelbi, A., 2009. Integrated
Research, 114 (2), 420429. model of preventive maintenance, quality control and
Lee, C.Y. and Chen, Z.L., 2000. Scheduling jobs and buffer sizing for unreliable and imperfect production
maintenance activities on parallel machines. Naval systems. European Journal of Operational Research,
Research Logistics, 47 (2), 145165. 47 (2), 389402.
Legnani, E., Cavalieri, S., and Ierace, S., 2009. A framework Rezg, N., Dellagi, S., and Chelbi, A., 2008. Joint optimal
for the configuration of after-sales service processes. inventory control and preventive maintenance policy.
Production Planning & Control, 20 (2), 113 124. European Journal of Operational Research, 46 (19),
Liao, L.-M. and Tsai, C.-H., 2009. Heuristic 53495365.
algorithms for two-machine flow shop with availability Rishel, T.D. and Christy, D.P., 1996. Incorporating main-
Downloaded by ["Queen's University Libraries, Kingston"] at 05:15 06 January 2015
constraints. Computers and Industrial Engineering, 56 (1), tenance activities into production planning; integration at
306311. the master schedule versus material requirements level.
Mirghani, M.A., 2001. A framework for costing planned International Journal of Production Research, 34 (2),
maintenance. Journal of Quality in Maintenance 421446.
Engineering, 7 (3), 170182. Ruiz, R., Garca-Daz, J.C., and Maroto, C., 2007.
Oyarbide-Zubillaga, A., Goti, A., and Sanchez, A., 2008. Considering scheduling and preventive maintenance in
Preventive maintenance optimization of multi-equip- the flow shop sequencing problem. Computers &
ment manufacturing systems by combining discrete Operations Research, 34 (11), 33143330.
event simulation and multi-objective evolutionary Tam, A.S.B., Chan, W.M., and Price, J.W.H., 2006. Optimal
maintenance intervals for a multi-component system.
algorithms. Production Planning & Control, 19 (4),
Production Planning & Control, 17 (8), 769779.
342355.
Taracki, H., et al., 2006. Maintenance outsourcing of a
Panesar, S.S. and Markeset, T., 2008. Industrial service
mulyi-process manufacturing system with multiple con-
innovation through improved contractual relationship:
tractors. IIE Transactions, 38 (1), 6778.
a case study in maintenance. Journal of Quality in
Weinstein, L. and Chung, C.H., 1999. Integrating main-
Maintenance Engineering, 14 (3), 290305.
tenance and production decisions in a hierarchical
Pistikopolos, E.N., et al., 2001. Interactions of maintenance
production planning environment. Computers &
and production planning for multipurpose process plants
Operations Research, 26 (1011), 10591074.
a system effectiveness approach. Industrial and Engineering
Yin, R.K., 1994. Case study research: design and
Chemistry Research, 40 (14), 31953207.
methods. 2nd ed. Thousand Oaks: Sage.
Portioli-Staudacher, A. and Tantardini M., 2008. Production
Yulan, J., Zuhua, J., and Wenrui, H., 2008. Multi-objective
and maintenance integrated management: proposal and
integrated optimization research on preventive mainte-
validation of a model in the case of outsourced nance planning and production scheduling for a single
maintenance. In: 10th international conference on the machine. International Journal of Advanced Manufacturing
modern information technology in the innovation Technology, 39 (910), 954964.