World Academy of Science, Engineering and Technology

International Journal of Economics and Management Engineering

Vol:6, No:8, 2012

Framework for Spare Inventory Management

Eman M. Wahba, Noha M. Galal, Khaled S. El-Kilany

Abstract—Spare parts inventory management is one of the major Thus, they are different from work-in-process (WIP) and
areas of inventory research. Analysis of recent literature showed that finished product inventories in inventory function and
an approach integrating spare parts classification, demand governing policies [4].
forecasting, and stock control policies is essential; however, adapting
These and other unusual aspects of spare parts inventories
this integrated approach is limited. This work presents an integrated
framework for spare part inventory management and an Excel based have led the authors to focus on the following work objective,
application developed for the implementation of the proposed which is to develop a general framework for spare inventory
framework. A multi-criteria analysis has been used for spare management, develop an Excel based application to support
classification. Forecasting of spare parts’ intermittent demand has this framework application, and to illustrate its effectiveness
International Science Index, Economics and Management Engineering Vol:6, No:8, 2012 waset.org/Publication/8567

been incorporated into the application using three different and practicality using a real-life case study.
forecasting models; namely, normal distribution, exponential
The remainder of this paper is organized as follows. Section
smoothing, and Croston method. The application is also capable of
running with different inventory control policies. To illustrate the 2 includes a brief review of relevant literature; in Sections 3
performance of the proposed framework and the developed and 4 the proposed framework of spare inventory management
application; the framework is applied to different items at a service is introduced along with the developed Excel based
organization. The results achieved are presented and possible areas application. In section 5, the application of the proposed
for future work are highlighted. framework to a real case study is presented and the
implementation results are discussed. Finally, conclusions are
Keywords—Demand forecasting, intermittent demand, inventory
drawn and areas for future research are suggested.
management, integrated approach, spare parts, spare part
classification
II. RELATED REVIEW
I. INTRODUCTION The variety of research in management of maintenance
inventories is very broad in scope [4]. To effectively manage
I N maintenance engineering, the cost of spare parts are
tremendous expenditure, and spare parts are very important
for equipment availability, they are significant resources in
inventories, three steps are needed; Spare Parts Classification,
Demand Forecasting and Inventory Control. Some authors
equipment maintenance [1]. The principal objective of any reviewed previous literature on spare inventory management
inventory management system is to achieve sufficient service such as [4] and [5].
level with minimum inventory investment and administrative The following sections present the review of the literature
costs [2]. Accordingly, the trade-off is clear: on one hand a related to the three requirements of an effective spare part
large number of spare parts ties up a large amount of capital; inventory management: Spare Parts Classification, Demand
while, on the other hand too little inventory may result in poor Forecasting and Inventory Control. Two approaches are
customer service or extremely costly emergency actions [3]. detectable in the literature: considering each of the three
In addition, spare part inventory management is a special requirements separately and the study of integrating these
case of general inventory management with some special requirements. Both approaches are covered in the review.
characteristics, spare parts are considered as very slow-moving A. Spare Parts Classification Review
parts with highly stochastic and erratic demands [1], derived
Large number of spare parts inventories makes the
by the different scheduled and unscheduled maintenance
management of inventory items difficult. As a result, parts are
operations.
commonly grouped and special stock control policies are
applied to each group. The most commonly used approach to
Eman M. Wahba, B. Sc., is a postgraduate student at the Department of classify inventory items is the ABC analysis. Traditionally, the
Industrial and Management Engineering; College of Engineering and
importance of a stock item is evaluated in terms of a single
Technology; Arab Academy for Science, Technology, and Maritime
Transport; AbuKir Campus, P.O. Box 1029, Alexandria, Egypt (phone: +203- criterion, the annual dollar usage of stock keeping units.
561-0755; fax: +203-562-2915; e-mail: [email protected]). However, other criteria may also play a significant role in
Noha M. Galal, Ph. D., is an Assistant Professor; Department of Industrial classifying stock keeping units such as lead time, criticality,
and Management Engineering; College of Engineering and Technology; Arab
Academy for Science, Technology, and Maritime Transport; AbuKir Campus, stock out consequences, demand rate, etc. In that case multiple
P.O. Box 1029, Alexandria, Egypt (e-mail: [email protected]). criteria ABC analysis for spare parts inventory is used. This
Khaled S. El-Kilany, Ph. D., is an Associate Professor and Chairman; has been addressed by different researchers [6]-[8] and various
Department of Industrial and Management Engineering; College of
Engineering and Technology; Arab Academy for Science, Technology, and
models have been developed to solve multiple criteria ABC
Maritime Transport; AbuKir Campus, P.O. Box 1029, Alexandria, Egypt (e- analysis as in [8]-[13].
mail: [email protected]).

International Scholarly and Scientific Research & Innovation 6(8) 2012 2220 scholar.waset.org/1307-6892/8567
 World Academy of Science, Engineering and Technology
International Journal of Economics and Management Engineering
Vol:6, No:8, 2012

It should be noted that special attention has also been given Moreover, an effective spare parts inventory control
to critical class of spare parts; where, demand for spare parts methodology motivated by a case study at a large oil refinery
can sometimes be classified into critical and non-critical has been presented [20]. The methodology starts with parts
demand depending on the criticality of the equipment for classification based on parts criticality, consumption rate,
production [14]. In addition, failure modes, effects and price, and lead time; followed by, demand forecasting using
criticality analysis (FMECA) has been used for evaluating different forecasting methods. Finally, an inventory model has
spare criticality [15]. been used as well as an optimization of the system based on
service level.
B. Demand Forecasting and Inventory Control Review
Finally, the integrated approach to spare parts inventory
As mentioned earlier, spare parts are characterized by management has also been addressed from a theoretical point
highly stochastic and erratic demands, which are derived by of view with no application to a real case study [21]. Again,
the different scheduled and unscheduled maintenance different methods of parts categorization have been presented
operations. This type of demand is generally referred to as and an inventory model has been introduced for optimization
intermittent demand and makes forecasting especially difficult. of stocking levels.
Traditional time series methods, such as moving average or This brief review of literature showed that few studies
single exponential smoothing are still the most commonly used
International Science Index, Economics and Management Engineering Vol:6, No:8, 2012 waset.org/Publication/8567

introduced the integrated approach for managing spare

in business practice [5]. Exponential smoothing places more inventories; hence, an integration of spare classification and
weight on the most recent data, resulting, in the case of demand forecasting with stock control policies is needed.
intermittent demand, in a series of estimates that are highest Furthermore, practical application of methods to real cases of
just after a demand occurrence and lowest just before demand spare management is often missing resulting in a gap between
occurs again [16]. research and practice.
Other methods have been developed specifically for
forecasting intermittent demand such as the one proposed by III. FRAMEWORK DEVELOPMENT
Croston, which forecast intermittent demand taking into
A general framework for planning and managing spare parts
account both demand size and the interval between demand
inventory is proposed with an objective of minimizing total
incidences. However, Croston method has been reported to
inventory cost and maximizing equipment availability taking
exhibit certain limitations and has been modified to improve
into consideration the desired level of safety. This framework
its accuracy of forecasts [16]. Furthermore, a modified
can be used for spare part inventory management for both
bootstrap method has been developed in order to take into
manufacturing and service organizations.
account autocorrelation, which proved to be an even better
The proposed framework presents an inventory control
method than Croston for forecasting spare parts’ intermittent
mechanism by integrating three steps. It starts with
demand [17].
classification of spare parts into groups, followed by spare
Finally, different approaches have been reportedly used to
parts demand forecasting and; finally, for each item class the
forecast spare parts demand with the objective of accurate
suitable ordering policy is defined based on the forecasted
determination of stock levels and taking into consideration the
demand. Each of these steps has a significant role for the
desired service level. These have been applied to real case
determination of the stock control method.
studies and showed that better forecasting of lead time demand
can result in reduction of stocking cost and stock out A. Spare Parts Classification
probability [3], [18] and [19]. A multi-criteria approach is adopted in the proposed
C. Integrated Approach Review framework for spare classification. Where, items are classified
based on the following three criteria: criticality, consumption
In order to create an effective inventory model for managing
rate, and inventory cost. These criteria are selected to address
spare parts, an integrated approach considering all
the proposed framework main objectives; safety, equipment
requirements for this effective inventory management can be
availability, and total inventory cost.
used. This approach should cover parts classification, demand
forecasting, and inventory control policies.
1) Criticality Classes
An effective method to manage spare part inventory, over
Maintenance work focuses on ensuring the continuous
three phases, has been introduced and implemented on a case
operation of equipment and maximum work efficiency. To
study of an Italian pasta producer [1]. Where, in the first
ensure this, spare parts are needed. Yet, the priority of having
phase, a FMECA analysis was performed to classify all items
different spare parts may vary. The lack of some spare parts
based on items criticality. Then, ABC method was used to
may cause equipment shutdown and other significant
evaluate spare parts consumption rate. Finally, a comparison
consequences. The spare part is defined as a critical one as it
between demand modeling techniques; Normal, Poisson, and
critical for equipment operation, expensive with a significant
Gamma distributions, was carried out.
variation in demand and long lead times. Accordingly,
availability of critical spare parts stock needs to be ensured
through an effective management method.

International Scholarly and Scientific Research & Innovation 6(8) 2012 2221 scholar.waset.org/1307-6892/8567
 World Academy of Science, Engineering and Technology
International Journal of Economics and Management Engineering
Vol:6, No:8, 2012

A systematic technique is used here to evaluate spare part 3) Inventory Cost Classes
criticality based on structuring the evaluation process into a Inventory management major conflict results mainly from
three-level hierarchy. At level 1, we define failure inventory total cost, usually we prefer to stock all needed spare
consequences of spare parts, i.e. the most reasonable parts however we face the problem of inventory costs. Total
consequences resulting from spare part failure or malfunction; inventory cost consists of three main components; holding
what could happen if the spare part is not available in stock, cost, ordering cost and shortage cost.
how it will affect the major concerns of the organization. The In this study shortage cost is ignored, since the shortage cost
definition of failure consequences is shown in Table I. is already considered in the criticality classification. The
TABLE I classification scheme used is ABC classification according to
DEFINITION OF FAILURE CONSEQUENCES Pareto principle.
Organization
Failure Consequences 4) Multi Criteria Classes
Concerns
The consideration of a number of criteria in classification
Safety A risk on people and environment safety; parts raises a significant problem of conflicting objectives; a spare
failure may result in death, disability or part may be considered as a critical item with high inventory
environmental damage that violates law or
International Science Index, Economics and Management Engineering Vol:6, No:8, 2012 waset.org/Publication/8567

regulations
cost and low consumption rate, which may lead in conflicts
while choosing the stock control method.
Equipment The effect on equipment operation consecutively Therefore, a management tool for multi-criteria cases is
Operation equipment availability; when a part fails it may needed. In this framework a matrix-based methodology is
cause total or partial equipment shutdown which
adopted, which is computationally simple, effective, and easy
result in less equipment availability.
to use.
Equipment A loss in money is always associated with First, a set of consecutive matrices are created; where, the
Downtime Cost equipment unavailability and maintenance work. number of matrices depends on the number of criteria
considered. In this paper 3 criteria are used as an example;
Products Quality (if For industrial and manufacturer organizations, inventory cost, consumption rate, and criticality class.
applicable) equipment is used to produce a product and if
one of this equipment is not working properly, According to the bi-matrix shown in Table II, items are first
this may affect products final quality. managed on the basis of consumption rate and inventory cost
regardless of the criticality level.
Customer In service organizations, customers’ satisfaction
Satisfaction (if is the organization main goal and concern. TABLE II
applicable) Unavailability of some equipment may affect CONSUMPTION RATE AND INVENTORY COST BI-MATRIX
directly the customer needs.
Consumption Rate

A Class B Class C Class

At the second level, our main target is to specify how the
A Class AA AB AC
Inventory

whole system works and the relation between spare parts and
Cost

equipment. Hence, we develop an equipment breakdown B Class BA BB BC

structure and maintenance plan structure. The criticality level C Class CA CB CC
is defined first as combined level then using priority level we
can define a critical class of each item.
Finally, at level three, we provide a qualitative measure for Then, as shown in Table III, the bi-matrix developed is
spare part criticality based on expert judgment. The number of combined with the criticality criteria. A combined class is now
criticality levels should be established depending on the size of defined for each stock item that includes the three criteria.
the system being considered to reflect the level of concern. Accordingly, we can select an inventory policy which better
We propose two categories of criticality levels; "High, suits the class characteristics as will be described in details in
Medium and Low" or "Catastrophic, High, Medium, Low and next section.
No effect". Accordingly, we first define a combined class for
each spare part then we assign priority level for each criterion TABLE III
COMBINED CLASSES
to evaluate the final critical class of the spare part.
Criticality Level
2) Consumption Rate Classes High Medium Low
Consumption of spare parts is an important criterion for AA AAH AAM AAL
Cost/Consumption

spare part classification. It is used to specify which spare parts

AB ABH ABM ABL
are more frequently used. The classification scheme used is
Inv.

ABC classification according to Pareto principle; where, class AC ACH ACM ACL
A includes spare parts with the most utilization percentage, BA BAH BAM BAL
class B includes spare with less utilization, and class C
BB BBH BBM BBL
includes spare with the least percentages of utilization.

International Scholarly and Scientific Research & Innovation 6(8) 2012 2222 scholar.waset.org/1307-6892/8567
 World Academy of Science, Engineering and Technology
International Journal of Economics and Management Engineering
Vol:6, No:8, 2012

BC BCH BCM BCL The output of these modules is the ordering policy of each
CA CAH CAM CAL
spare part. Furthermore, it can be easily configured by users to
match their current business processes and requirements. This
CB CBH CBM CBL
section presents a step by step guide through the developed
CC CCH CCM CCL computer model showing its main and special features of its
constituting modules.
The second step is assigning a priority level for each A. Initial Data Setup
criterion, priority levels are chosen on system basis. In the
Users start using this application by adding the initial data
present study we propose a general structure for priority levels;
required as shown in Fig. 1.
criticality has the highest priority level. The class including
spare parts with high criticality level must be physically
present at the warehouse and controlled using safety stocks;
regardless of the other two criteria; due to the catastrophic
effects that can happen if they were not available.
The rest of classes will be managed on class basis using the
International Science Index, Economics and Management Engineering Vol:6, No:8, 2012 waset.org/Publication/8567

priority levels defined. For example if second priority after

safety is the consumption rate, items that belong to combined
class "ACL", i.e. having high inventory cost, low consumption,
and low criticality level are managed using just-in-time policy.
However, items in class "CAM", i.e. having low inventory
cost, high consumption rate, and medium criticality are
managed using continuous review inventory policy (safety
stock policy).
Fig. 1 Initial spare parts data entry screen
B. Demand Forecasting and Inventory Control
After the definition of all combined classes and the selection After setting initial data, users can define what criteria and
of a suitable inventory policy for each class, we need to methods are to be used, as shown in Fig. 2. First they select the
calculate the parameters of the selected inventory policy; such classification scheme, whether single or multi-criteria and the
as, the reorder point and order quantity for each stock item to type of criteria considered. Then they select the method for
achieve a predefined service level. forecasting demand of spare parts. Three methods are
Economic Order Quantity (EOQ) is usually used to available; normal distribution, single exponential smoothing
determine the order quantity and is evaluated average and Croston method. Finally, they can specify which reports
forecasted annual demand. In contrast, calculating the reorder are needed.
point requires estimates of the lead time demand (LTD). As
mentioned earlier in the literature review section, different
forecasting methods have been applied to determine the lead
time demand. In this work, both probability distributions and
forecast techniques are used, in order to arrive at a robust
framework that can be implemented in many situations and
real case studies.
The normal distribution is selected for modeling spare parts
demand as a probability distribution, since it performs better in
business practice, provides more accurate results, and is
simpler to use. In addition, exponential smoothing and Croston
method are selected as two competing forecasting methods.

IV. COMPUTER MODEL DEVELOPMENT

An Excel based application for spare part inventory Fig. 2 Main screen
management has been developed. It provides an effective and B. Spare Parts Classification Module
user friendly tool for applying the proposed framework for
1) Criticality Evaluation
small sized organizations. The main components of the We built this evaluation process based on five questions
application are a startup screen for data input and three main concerning the effect of spare unavailability on safety, system
modules for basic operations; spare part classification, operation, downtime cost, products’ quality, and customer
inventory management, and reporting module. satisfaction; as shown in Fig. 3.

International Scholarly and Scientific Research & Innovation 6(8) 2012 2223 scholar.waset.org/1307-6892/8567
 World Academy of Science, Engineering and Technology
International Journal of Economics and Management Engineering
Vol:6, No:8, 2012

3) Multi-Criteria Analysis
The multi-criteria analysis module shown in Fig. 5 is
responsible of assigning priority levels for each criterion
through any of the following alternatives:
a) Users can customize priority level for each criterion based
on maintenance requirements and objectives.
b) Users can use the system default settings, which is
criticality, consumption rate, then total inventory cost.
Using the defined priority levels the module will calculate the
combined class and consequently the final class.
Fig. 3 Criticality evaluation screen

This module allows users to:

a) Select which questions are applicable. For example a
product quality may be inapplicable in a service
organization. Furthermore, if the available questions are
International Science Index, Economics and Management Engineering Vol:6, No:8, 2012 waset.org/Publication/8567

inconvenient, the user can add/replace available ones (a

maximum of five questions can be used).
b) Assign priority levels for questions. Users can set the
priority levels according to their needs.
c) Choose the criticality level category. The user can choose Fig. 5 Multi-criteria analysis module
between "High, Medium, and Low" or "Catastrophic,
C. Inventory Management and Demand Forecasting Module
High, Medium, Low, and No effect".
d) Select if criticality will be evaluated for each part or for In this module, the inventory policy is selected and the
each class of parts if needed. reorder point, EOQ, and safety stock level are calculated.
e) Evaluate for each part the combined criticality class;
consequently, the module will determine its criticality 1) Inventory Policy Selection
class using the defined priority levels. Inventory policy can be selected either automatically or
f) Generate a summary report of the number of stock items manually. Based on the classes defined for each item in
in each class and its percentage out of the total number of previous step, the model can automatically select the inventory
stock items. policy; class A and B are managed using continuous review
2) Consumption Rate and Total Inventory Cost Calculations inventory policy; however, for class C, just-in-time policy is
The consumption rate and total inventory cost modules are applied. Furthermore, users can select the inventory policy
carried out in a similar way, using ABC method as shown in according to their own judgment regardless of the item class.
Fig. 4.
2) Forecasting Demand and LTD Modeling
This module is responsible of forecasting demand, modeling
LTD, and consequently determining the parameters of the
inventory policy; reorder point and order quantity for each
item. Of course these values are determined at a specific
service level as shown in Fig. 6.

Fig. 4 Consumption rate calculation screen

However, sometimes the results from the ABC method can

be inconvenient for some users; therefore, we added a
sensitivity analysis capability, which will update the results
according to an acceptable range, defined by users, for each
class. Same as criticality process, the user can generate a Fig. 6 ROP calculation screen
summary report of the number of stock items in each class and
its percentage out of the total number of stock items.

International Scholarly and Scientific Research & Innovation 6(8) 2012 2224 scholar.waset.org/1307-6892/8567
 World Academy of Science, Engineering and Technology
International Journal of Economics and Management Engineering
Vol:6, No:8, 2012

A. Reporting Module The data collection includes general information about the
This module summarizes all results obtained from the other organization, equipment function and hierarchy, maintenance
modules. Moreover, this module provides a mean to evaluate operations, stock items (number of warehouses, demand,
the difference between the three methods used for demand prices, lead time, and current quantities), inventory
forecasting by calculating average inventory level and the management methods, and processes. Interviews with
average inventory value for each item. responsible persons are conducted to gather this information.
The average inventory level is calculated based on the The service organization under study consists of a large
average minimum and maximum level of inventory. The complex that includes a considerable number of assets and
aspired minimum level is the safety stock; while, the aspired systems used to operate all aspects of the building. This
maximum level is the sum of the safety stock and the EOQ. organization has one main warehouse with 6343 items;
The same calculation applies to the inventory values as including all materials and spare parts needed for maintenance
shown in Fig. 7. and repair. This warehouse is divided into 6 storage areas
according to maintenance systems, e.g. mechanical stock, heat
ventilation and air conditioning (HVAC) stock, electrical
supplies…etc. Currently, an enterprise asset management
system (EAM), a management system for maintenance and
International Science Index, Economics and Management Engineering Vol:6, No:8, 2012 waset.org/Publication/8567

spare parts, is used. It provides the tool to manage equipment

maintenance and to keep track of warehouses main data
(transactions, current quantities, reservations...etc.) but it is not
specific for inventory control. Therefore, available stores are
managed based on expert judgment whether for spare part
classifications or demand forecasting.
In addition, parts needed are ordered on demand, i.e. no
inventory policy is applied which results in high stock levels
and thus high inventory cost. Accordingly, the main objective
of this study is to define how much stock is needed for each
Fig. 7 Inventory value report screen item taking into consideration different criteria such as item
criticality, price, demand rate...etc. to ensure low stock levels
Furthermore, the reporting module calculates total savings associated with high level of equipment operation.
achieved using the proposed methods in comparison to current Demand information is collected on monthly basis for 1
inventory value as shown in Fig. 8. year (from January 2011 to January 2012) so a total of 12
periods is available. Demand set is not specified whether it is
due to preventive or corrective maintenance. We exclude parts
with negative or zero annual demand. We classify stock items
in two demand categories, spare parts and consumables. Spare
parts are needed for maintenance or repair therefore they are
characterized with an intermittent demand with a large
proportion of zero values. While consumables are used more
frequently to maintain a high level of equipment operation
which resulted in a more regular demand with less proportion
of zero values. In the present study, both spare parts and
consumables are considered while forecasting demand.
According to prices, parts with average price zero are not
owned by the maintenance department. Therefore, they are not
Fig. 8 Savings report screen associated with any holding cost, consequently these items are
not considered while evaluating the inventory model.
V. CASE STUDY Interviews are conducted with maintenance engineers,
inventory specialists and top management to define their main
A. Problem Definition and Data Collection
concerns, goals and available resources. We find out that
In order to evaluate how the framework can be taken into safety and equipment operation are the major two concerns of
practice, test the developed computer model, and illustrate its the organization. They give first priority to safety and require a
performance; the proposed framework is applied to a case service level of 95% for equipment operation. Additional
study at a service organization. We investigate the inventory interviews are also organized with responsible maintenance
management system of this service organization, gather
engineers to determine the criticality level of stock items.
historical demand data and their corresponding facility
They choose a criticality level category of "High, Medium
maintenance information then use these data in the
implementation of the proposed framework using the and Low".
developed spread sheet based application.

International Scholarly and Scientific Research & Innovation 6(8) 2012 2225 scholar.waset.org/1307-6892/8567
 World Academy of Science, Engineering and Technology
International Journal of Economics and Management Engineering
Vol:6, No:8, 2012

B. Findings Summary The results are compared to the current inventory value and
Using the collected data and the developed computer model, total savings are calculated for each store. Fig. 11 clearly
spare classification processes and demand forecasting are shows that all proposed methods provide better results than the
carried out; consequently, we calculate the reorder point and current system. Expect for, using Croston method in store 6,
EOQ for each stock item. which does not result in any savings; however, on the contrary
However, we still cannot judge the overall effectiveness of negative results are achieved.
the proposed framework. According to decision makers, they
still do not have a clear picture on which forecasting method is
better. Hence, performance measurement is carried out using
the reporting module. For each store we calculate optimum
inventory level and its value. Furthermore, total savings
achieved from the adoption of the proposed framework are
calculated.
Fig. 9 shows that using the exponential smoothing method
results in minimum inventory value in comparison to Croston
International Science Index, Economics and Management Engineering Vol:6, No:8, 2012 waset.org/Publication/8567

method and Normal distribution except for store 2.

Fig. 11 Total savings

Further investigation of the results obtained for store 6

shows that it contains more consumables than spare parts. In
other words, it has a more stable demand than that of the
remaining stores. This is why we obtain the least savings when
using the exponential smoothing method and even obtained
negative results when using the Croston method. This proves
that the proposed framework is more suitable to use for
managing spare parts inventory having lumpy demand.
Furthermore, single exponential smoothing method provides
Fig. 9 Stores Inventory Value
the least inventory value as well as the highest cost savings
On the other hand, Fig. 10 shows that for almost all stores, with respect to other methods. Savings vary from 46% for
Normal distribution provided the least inventory level. store 6 to approximately 90% for store 1.
Furthermore, we observe that the majority of stores have Finally, Croston method, despite its ability to provide more
current inventory levels less than the forecasted levels, estimates of demand per period, has performed the worst of all
however their inventory value is much greater, this is resulting methods, the inventory level was overestimated and the
from stocking non critical and expensive spare parts. This inventory value is the highest among the models used.
result reflects the existing gap between spare part classification
process and inventory control. Therefore, we conclude that VI. CONCLUSION
using the integrated perspective stresses the relation between
Spare part inventory management is a vivid field of research
spare part classification and inventory management, resulting
due to the importance of the items under control, their
in less inventory value for the same service level.
influence on the operations and the difference in demand
structure from other stocking items. From the conducted
literature review it can be concluded that an integrated
approach is effective for managing inventory of spare parts.
Few works have been presented in this subject area and
practical case studies are hardly presented. The objective of
the current work is to fill in this gap by proposing an
integrated framework for spare part inventory management and
providing a real life case study implementation.
The proposed framework comprises spare part
classification, demand forecasting and inventory management.
In order to facilitate and automate the implementation of the
proposed framework, we develop a spread sheet based
Fig. 10 Stores Inventory Level application for spare part inventory management.

International Scholarly and Scientific Research & Innovation 6(8) 2012 2226 scholar.waset.org/1307-6892/8567
 World Academy of Science, Engineering and Technology
International Journal of Economics and Management Engineering
Vol:6, No:8, 2012

It provides an easy to use tool for inventory control; our [13] A. Hadi-Vencheh and A. Mohamadghasemi, “A fuzzy AHP-DEA
approach for multiple criteria ABC inventory classification,” Expert
main target development is to create a generic application Systems with Applications, vol. 38, Issue 4, pp. 3346-3352, April 2011.
suitable for a wide range of case studies. [14] P. Long Chang, Y. Chou, and M. Guang Huang, “A (r; r;Q) inventory
In order to verify and validate the proposed framework, it is model for spare parts involving equipment criticality,” International
applied to a case study at a service organization using real Journal of Production Economics, vol. 97, Issue 1, pp. 66-74, 18 July
2005.
historical data of 6343 items; including all materials and spare [15] S. Chin Yang and Z. Du, “Criticality evaluation for spare parts initial
needed for maintenance and repair. The conclusion drawn provisioning,” in Reliability and Maintainability, 2004 Annual
from the implementation results is that an integrated approach Symposium, RAMS, 26-29 Jan. 2004, pp. 507- 513.
[16] A. A Syntetos and J. E Boylan, “On the bias of intermittent demand
leads to lower inventory levels for the required service level. estimate,” International Journal of Production Economics, vol. 71,
In addition we may conclude that Exponential Smoothing Issues 1–3, pp. 457-466, 6 May 2001.
technique for demand forecasting performs better than the [17] T.R. Willemain, C.N. Smart, and H.F. Schwarz, “A new approach to
forecasting intermittent demand for service parts inventories,”
other two methods; normal distribution and Croston method. International Journal of Forecasting, vol. 20, Issue 3, pp. 375-387,
This present study and other integrated approaches July–September 2004.
presented in literature, did not take into consideration the type [18] S.G. Li, and X. Kuo, “The inventory management system for automobile
spare parts in a central warehouse,” Expert Systems with Applications,
of maintenance policy applied whether it is scheduled or
vol. 34, Issue 2, pp. 1144-1153, February 2008.
International Science Index, Economics and Management Engineering Vol:6, No:8, 2012 waset.org/Publication/8567

unscheduled. Furthermore, integration between spare part [19] Z. Hua, and B. Zhang, “A hybrid support vector machines and logistic
management and maintenance intervals may be studied. Since regression approach for forecasting intermittent demand of spare parts,”
maintenance operations directly affect spare parts inventory Applied Mathematics and Computation, vol. 181, Issue 2,pp. 1035-
1048, October 2006.
management, further study is needed in this field and an [20] E. Porras, and R. Dekker, “An inventory control system for spare parts
attention to maintenance intervals while implementing spare at a refinery: An empirical comparison of different re-order point
part classification, demand forecasting and inventory control is methods,” European Journal of Operational Research, vol. 184, pp. 101-
132, January 2008.
needed. [21] K. Yang, and X. Niu, “Research on the spare parts inventory,”
Industrial Engineering and Engineering Management, 2009. IE&EM
REFERENCES '09. 16th International Conference on, pp. 1018-1021, 21-23 Oct. 2009.
[1] Y. Li, B. Yang, H. Hu, and I. Chang, “A simulation-optimization
approach to spare parts allocation based on decision-maker's
satisfaction,” in Reliability, Maintainability and Safety, 2009. ICRMS
2009. 8th International Conference on, pp. 576-581, 20-24 July 2009.
[2] M. Bevilacqua, F.E. Ciarapica, and G. Giacchetta,“Spare parts
inventory control for the maintenance of productive plants,”" in
Industrial Engineering and Engineering Management, pp.1380-1384,
8-11 Dec. 2008.
[3] K. Aronis, I. Magou, R. Dekker, and G. Tagaras, “Inventory control of
spare parts using a Bayesian approach: A case study,” European
Journal of Operational Research, vol. 154, Issue 3, pp. 730-739, May
2004.
[4] W.J. Kennedy, J. Wayne Patterson, and L. D. Fredendall, “An overview
of recent literature on spare parts inventories,” International Journal of
Production Economics, vol. 76, pp. 201-215, March 2002.
[5] A. Bacchetti and N. Saccani, “Spare parts classification and demand
forecasting for stock control: Investigating the gap between research and
practic,” Omega, vol. 40, Issue 6, pp. 722-737, December 2012.
[6] L. Ye Chen and K.W., Si-feng Liu; , “A comparative study on
multicriteria ABC analysis in inventory management,” in Systems, Man
and Cybernetics, 2008. SMC 2008. IEEE International Conference on,
pp. 3280-3285, 12-15 Oct. 2008.
[7] I. Liiv, “Inventory classification enhancement with demand
associations,” in Service Operations and Logistics, and Informatics,
IEEE International Conference, SOLI, 2006, pp. 18-22, 21-23..
[8] M. Yu, “Multi-criteria ABC analysis using artificial-intelligence-based
classification techniques,” Expert Systems with Applications, vol. 38,
Issue 4, pp. 3416-3421, April 2011.
[9] W. Lung Ng, “A simple classifier for multiple criteria ABC analysis,”
European Journal of Operational Research, vol. 177, Issue 1, pp. 344-
353, 16 February 2007.
[10] P. Zhou and L. Fan, “A note on multi-criteria ABC inventory
classification using weighted linear optimization,” European Journal of
Operational Research, vol. 182, Issue 3, pp. 1488-1491, 1 November
2007.
[11] F. Y. Partovi and M. Anandarajan, “Classifying inventory using an
artificial neural network approach,” Computers & Industrial
Engineering, vol. 41, Issue 4, pp. 389-404, February 2002.
[12] C. Wu Chu, G. Shuh Liang and C. Tseng Liao, “Controlling inventory
by combining ABC analysis and fuzzy classification,” Computers &
Industrial Engineering, vol. 55, Issue 4, pp. 841-851, November 2008.

International Scholarly and Scientific Research & Innovation 6(8) 2012 2227 scholar.waset.org/1307-6892/8567