Professional Documents
Culture Documents
Access to this document was granted through an Emerald subscription provided by emerald-
srm:184293 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald
for Authors service information about how to choose which publication to write for and submission
guidelines are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as
well as providing an extensive range of online products and additional customer resources and
services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the
Committee on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for
digital archive preservation.
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
JQME
21,4
Multi-criteria classification of
spare parts inventories – a web
based approach
456 S.P. Sarmah and U.C. Moharana
Received 25 April 2012
Department of Industrial and Systems Engineering,
Revised 10 August 2013 Indian Institute of Technology Kharagpur, Kharagpur, India
23 November 2014
Accepted 30 June 2015
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
Abstract
Purpose – The purpose of this paper is to present a fuzzy-rule-based model to classify spare parts
inventories considering multiple criteria for better management of maintenance activities to overcome
production down situation.
Design/methodology/approach – Fuzzy-rule-based approach for multi-criteria decision making is
used to classify the spare parts inventories. Total cost is computed for each group considering suitable
inventory policies and compared with other existing models.
Findings – Fuzzy-rule-based multi-criteria classification model provides better results as compared to
aggregate scoring and traditional ABC classification. This model offers the flexibility for inventory
management experts to provide their subjective inputs.
Practical implications – The web-based model developed in this paper can be implemented in
various industries such as manufacturing, chemical plants, and mining, etc., which deal with large
number of spares. This method classifies the spares into three categories A, B and C considering
multiple criteria and relationships among those criteria. The framework is flexible enough to add
additional criteria and to modify fuzzy-rule-base at any point of time by the decision makers. This
model can be easily integrated to any customized Enterprise Resource Planning applications.
Originality/value – The value of this paper is in applying Fuzzy-rule-based approach for Multi-criteria
Inventory Classification of spare parts. This rule-based approach considering multiple criteria is not very
common in classification of spare parts inventories. Total cost comparison is made to compare the
performance of proposed model with the traditional classifications and the result shows that proposed
fuzzy-rule-based classification approach performs better than the traditional ABC and gives almost the
same cost as aggregate scoring model. Hence, this method is valid and adds a new value to spare parts
classification for better management decisions.
Keywords Fuzzy logic, Multi-criteria decision making, Inventory classification, Rule base, Spare parts
Paper type Research paper
1. Introduction
Efficient and effective spare parts inventory management plays a significant role in
maintaining a firm’s competitive advantage. The major focus of spare parts inventory
management practices is to have the required spares to be ready at right time, in right
place with minimum cost. Most of the companies store thousands of spares in
inventory and often they are not properly categorized on the basis of management
attention. It may cause severe production down situation when critical spares are
wrongly classified. Considering the importance and inspection levels, there may be a
chance that only a very few spares may require special attention. Therefore, an
accurate and robust classification of items is required to manage the large number of
Journal of Quality in Maintenance
Engineering spares. Rapid use of information technology and implementation of Enterprise
Vol. 21 No. 4, 2015
pp. 456-477
Resource Planning (ERP) in industry has provided us the flexibility of differentiating
© Emerald Group Publishing Limited
1355-2511
the spares. Practically, it is very difficult and also economically infeasible to apply
DOI 10.1108/JQME-04-2012-0017 different control schemes to individual spare part. It is necessary to group the spares
into different categories and applies different ordering policies according to the Classification
importance of each category. It will reduce the amount of effort necessary for managing of spare parts
the stock of the spares that are falling in the same group and at the same time, number
of spares that require more management attention will also be reduced.
inventories
One of the most widely used practices for classification of items in industry is
ABC analysis, which is based on Pareto’s principle. This is a classification that requires
single criterion, annual usage value (cost), i.e. product of annual use of the item and 457
its price. Few items, which contribute very large amount of annual usage value,
are considered as A class items. The large numbers of item, which contribute very
less amount of annual usage value, are considered as C class and B class items fall in
between these two categories. However, there are certain limitations of this type
of classification since it does not consider some crucial dimensions of spare items
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
2. Review of literature
Research articles related to spare parts inventory classification is less compared to spare
parts demand forecasting and inventory control. In this section, we have provided a brief
review of literature related to different approaches proposed in spare parts inventory
classification.
JQME 2.1 Spare parts classification using single criterion
21,4 Few researchers have developed single criterion spare parts classification models,
which are based on annual consumption value, demand variance or annual movement
of spare parts. Gelders and Van Looy (1978) have carried out a study on a large
petrochemical plant where the warehouse contains huge numbers of spare parts and
equipment. They suggested ABC classification using the demand transactions per year
458 and suitable inventory control model for different classes of spare parts. Williams
(1984) proposed a method for classification of spare parts demand into smooth,
slow moving or sporadic by partitioning the variance of demand during the lead-time.
Later, the study was extended to find out demand distribution of each category of
spares and to apply forecasting methods and stock control policies. A demand
classification model was developed by Syntetos et al. (2009), which helped to determine
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
the stock control-related opportunities for increasing service levels and reducing costs
of spare parts. A detailed study was carried out for demand occurrence frequency over
six months along with number of orders to find out required stock level of spare parts.
perishability of items and cost of storing of raw materials were considered for the case
study. They also used ANOVA and simulation study to find out the effectiveness
and comparison of proposed model with traditional ABC classification. Tsai and
Yeh (2008) used particle swarm optimization approach for inventory classification.
In this classification, inventory items are classified based on a specific objective or
multiple objectives, such as minimizing costs, maximizing inventory turnover ratios,
and maximizing inventory correlation. This approach determines the best number
of inventory classes and how items should be categorized for the desired objectives
at the same time. Chen (2011) proposed a peer-estimation approach for MCIC which
determines two common sets of criteria weights and aggregates resulting two
performance scores in the most favorable and least favorable senses for each item
without any subjectivity. A comparison was made for the proposed approach with
some previous methods and proved that the proposed model could provide a more
reasonable and comprehensive performance index for MCIC.
3. Methodology
In this paper, a new model called FRMIC is developed to classify spare parts
considering the association between criteria. The basic concept of fuzzy If-then rule is
used for the classification of spare parts considering different input parameters and
ABC output parameter. Here, a brief description of aggregate scoring model is also
incorporated which follows multi-criteria decision-making approach for comparison
purpose with our proposed model (reader can refer Flores et al., 1992).
3.1 Fuzzy-rule-based-model
Before presenting our model, a brief description of fuzzy-rule-based model
(Zadeh, 1965) is given here for easy understanding to the reader. A fuzzy inference
system is also known by other names such as fuzzy-rule-based system, fuzzy model,
fuzzy associative memory, and simply fuzzy system. The basic structure of a fuzzy
inference system consists of three conceptual components: a rule base, which contains
a selection of fuzzy rules; a database that defines the membership functions used in
the fuzzy rules; and a reasoning mechanism, which performs the inference procedure
upon the rules and given facts to derive a reasonable output or conclusion. A fuzzy
inference system implements a non-linear mapping from its input space to output
space. This mapping is accomplished by a number of If-then rules, each of which
describes the local behavior of the mapping. In particular, the antecedent defines
a fuzzy region in the input space, while the consequent specifies the output in the
JQME fuzzy region. A method of defuzzification is needed to extract a crisp value that best
21,4 represents the fuzzy set of the output. The block diagram in Figure 1 shows a fuzzy
inference system.
The aggregator combines the fuzzy rules in the rule-based system to yield fuzzy
output, and is converted into a crisp output by a defuzzifier. In this paper, we have
adopted the procedure from Mamdani model, which is described below.
460 Mamdani fuzzy model. The Mamdani (Mamdani and Assilian, 1975) fuzzy
interference system uses min and max for T-norm and T co-norm operators,
respectively. The rules in this system are of the following form:
If x is A1 and y is B1 then z ¼ C1
If x is A2 and y is B2 then z ¼ C2
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
The crisp output is obtained as the centroid of the area of Z and is determined by the
following method:
R
mF ðzÞzdz
Z cg ¼ Rz
z mF ðzÞdz
where, μF ( z) is the aggregated output membership function.
We have shown here a simple two-input one-output problem (see Figures 2 and 3)
that includes three rules:
Rule 1: If x is A3 or y is B1 then z is C1
Rule 2: If x is A2 and y is B2 then z is C 2
Rule 3: If x is A1 then z is C 3
The defuzzified output value as shown in Figure 4 is calculated as follows:
ð0 þ 10þ 20Þ 0:1 þ ð30 þ 40 þ 50þ 60Þ 0:2 þ ð70 þ 80 þ 90 þ 100Þ 0:5
COG ¼
0:1 þ 0:1 þ 0:1 þ 0:2 þ 0:2 þ 0:2 þ 0:2 þ 0:5 þ 0:5 þ 0:5 þ 0:5
¼ 67:4
Figures 2-4 shows the complete procedure to obtain the crisp output of the above rules
using max-min composition.
Inference Mechanism
Defuzzification
Fuzzification
Figure 1. Rule-Base
Fuzzy-rule-based
system
1 1 1 Classification
A3 B1 C1 C2 C3 of spare parts
0.1 OR 0.1
0.0
(max)
inventories
0 x1 X 0 y1 Y 0 Z
Rule 1: IF x is A3 (0.0) OR y is B 1 (0.1) THEN z is C 1 (0.1)
1 1 1 461
0.7
C1 C2 C3
A2 0.2 B2 AND 0.2
(min)
0 x1 X 0 y1 Y 0 Z
Rule 2: IF x is A 2 (0.2) AND y is B 2 (0.7) THEN z is C 2 (0.2)
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
1 1
A1 0.5 0.5 C 1 C2 C3
Figure 2.
Pictorial
0 x1 X 0 Z representation
of rules
Rule 3: IF x is A 1 (0.5) THEN z is C 3 (0.5)
1 1 1
C1 C2 C3
0.5 0.5
0.2
Figure 3.
0.1 0.2 0.1 Pictorial
0 Z 0 Z 0 Z 0 Z representation
of defuzzification
z is C 1 (0.1) z is C 2 (0.2) z is C 3 (0.5) ∑ of output
Degree of
Membership ( )
1.0
0.8
0.6
0.4
0.2
Figure 4.
0.0 Determination of
0 10 20 30 40 50 60 70 80 90 100 center of gravity
Z
of output
67.4
In this paper, we have assigned equal weights to each of the criteria and the
aggregate score is determined by the addition of all criteria-specific scores.
The aggregate score shares the importance of each criteria and hence this score can
be used for MCIC for spare parts. The following steps give a brief description of
Aggregate Scoring Model:
CV max CV i
Consumption V alue Score ðS CV Þ ¼
CV max CV min
JQME U P max U P i
U nit price Score ðS U P Þ ¼
21,4 U P max U P min
LT max LT i
Leadtime Score ðS LT Þ ¼
LT max LT min
462
where CVmax is the maximum consumption value among all spares; CVmin the
minimum consumption value among all spares; CVi the consumption value of ith
spare; UPmax the maximum unit price among all spares; UPmin the minimum unit
price among all spares; UPi the unit price of ith spare; LTmax the maximum lead-time
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
among all spares; LTmin the minimum lead-time among all spares; and LTi the
lead-time of ith spare.
We do not need to scale commonality and criticality as these are already in 0-1
scales. Hence aggregate score of a spare is the sum of all five scores and can be
calculated as: S ¼ SCV + SUP + SLT + SCR + SCM. Now, the spares can be sorted as per
the aggregate score in descending order and 5 percent of items which have maximum
scores can be extracted as A or high important class, next 15 percent of items can be
considered as B or medium important class and rest of the items can be classified as
C or less important item.
the spare and the time of supply of spares. Lead-time can be divided into internal and
external lead-time. Internal lead-time is the time taken to prepare the requisition
document, selection of suppliers and placing the purchase orders to them. In this case,
there is a need for an efficient procurement management system to reduce the internal
lead-time. On the other hand, external lead-time is fully dependent on suppliers. In this
case, one should have a better method of supplier selection, supplier performance review,
periodic audit and co-ordination with suppliers to reduce the external lead-time. Also the
length of the lead-time and its variability could be important in maintaining an adequate
supply of an item without excessive costs (Flores and Whybark, 1986). The length of the
lead-time is important since it directly determines the stock levels of items with unknown
demand and dictates the response time to a crisis. The variability also affects the amount
of safety stock required to provide the level of service desired.
4.4 Commonality
Commonality is another measure to determine the number of equipment where the
same spare parts is used. Therefore, it is important to give more attention to the spare,
which shows high commonality factor, and these spares should be grouped to
important class for better management purpose. Commonality should be an important
criterion because if a part is common to majority of equipment, then there is a chance of
severe production down situation due to the shortage of this part. One spare part can be
very critical to some equipment and at the same time, it may be desirable for others
depending upon equipment criticality. Hence if a spare is used in many critical types of
equipment, commonality has greater effect toward the classification. Even if total
consumption considers the consumption of common parts together, both the criteria,
consumption value and commonality have different level of importance. One spare part
can have low commonality and high consumption value and it means if this spare
becomes stock out, only few equipment will be down. In the reverse case, one spare part
can have high commonality and low consumption value. Therefore, use of common
spare parts can be beneficial in terms of risk sharing and substantial repair or
maintenance time reduction by using shared stocks compared to the use of separate
stocks. If the same spare part is used with various equipment, it is beneficial to procure
the spare from a supplier with larger volumes with discounting policy.
Commonality index (CIi) as suggested by Wazed et al. (2009) is determined as
follows and the value of the index varies from 0 to 1:
N max N i N i N min
CI i ¼ 1 ¼
N max N min N max N min
JQME where Nmax is the maximum number of equipment sharing a particular spare; Nmin is
21,4 the minimum number of equipment sharing a particular spare; and Ni is the number of
different equipment that sharing the spare i.
4.5 Criticality
The criticality of a spare is another very important factor to be considered for
464 specifying service levels. DIC classification is widely used for spare parts as desirable,
important, and critical from the viewpoint of their functional necessity in production
and maintenance operations. Critical spare parts include all spares, which, if not in
stock, could result in huge losses due to non-availability of the equipment needing the
spare. Important spare parts are those for which stock-outs could result in moderate
losses. In this case, the equipment may be operable with some difficulty but cannot be
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
used for long periods without the spare. Non-availability of desirable spares will cause
only minor disruptions but may lead to more serious operational problems in the long
run (Gajpal et al., 1994). In our study, we have scaled the criticality index like desirable
class with range 0.0-0.04, important class with range 0.2-0.8 and critical class with value
0.6-1.0 for designing the rule base.
their different logins at different locations in the organization. This model can be easily
integrated to any of the online inventory management systems. The dynamic results of
the multi-criteria classification of spares can be obtained based on daily transaction
data. The architecture of FRMIC model is shown in Figure 5, which has the following
components:
(1) Gathering inventory record: this process collects all sorts of information for the
spare parts, which are required for FRMIC model like lead-time, unit price,
consumption values, criticality, etc. for each spare item from the database.
Gathering of information can be done on periodic basis like monthly, quarterly
or yearly depending upon the requirements from inventory control unit.
(2) Data cleaning and quantification of qualitative factors: it is very important to
clean the junk records before classification of items; otherwise it will lead to
wrong output. One should not consider an item, which has annual, zero demand
and zero lead-time. These items are always considered as junk record. Structured
Query Language is normally used for filtering/cleaning the data from inventory
database, which is a very common language for any of the database management
systems. Since we have used fuzzy logic concepts for classification, we need to
Generated Reports
Figure 5.
Traditional ABC and Schematic
Comparison ABC,
Aggregate scoring FRMIC Aggregate scoring
Fuzzy rules architecture of web-
Classification Process report
and FRMIC based FRMIC model
JQME quantify the qualitative factors and vice versa wherever it is necessary. In our
21,4 case, we have converted the criticality criteria to numeric value for defining the
membership function.
(3) Traditional ABC and aggregate scoring classification process: the traditional ABC
classification can be done using the standard procedures such as A class items
(high-value items) where 15-20 percent of the items that account for 75-80 percent
466 of the total annual inventory value, B class items (medium-value items) where
30-40 percent of the items that account for approximately 15 percent of the total
annual inventory value and C class items (low-value items) where 40-50 percent of
the items that account for 10-15 percent of the total annual inventory value.
Similarly, aggregate scoring can be done using the steps described in Section 3.2.
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
(4) User interface to modify/delete rule: in this process, experts can modify or delete
the fuzzy rules by using the “Modify/Delete Rule” page. This interface allows a
user-friendly way of creating rules, modifying and deleting the existing/
duplicate rules. This process is used when system finds any association
between criteria.
(5) FRMIC process: this process takes the crisp input criteria values and does
fuzzification of each criterion, inference mechanism with rule base. For each
spare, few rules out of 243 rules for the five criteria are selected for getting the
output membership values.
(6) Report generation: this process generates three reports, namely: fuzzy rules
report, FRMIC report and comparison between traditional ABC, aggregate
scoring and FRMIC classification report (Figure 5), which are accessible to both
decision makers and to the experts. The expert can check the correctness of
the rules and modify as per the associability between criteria. Experts can
also add new criteria and generate new rules depending upon the actual
requirement for a system.
6. Case study
The developed model is applied for 1,360 spare items of a reputed public sector open
cast mining company situated at southern part of India. The company handles nearly
150,000 of items and among those items, 1,360 spare parts are selected for classification
based on availability of information required for MCIC. Five, criteria namely:
consumption value, unit price, lead-time, commonality and criticality are considered
here for classification. Unlike in other literature, we have introduced a new criterion
known as commonality, which is determined by the commonality index, and its
importance is already explained in Section 4. All the criteria are positively related to the
score of the inventory items. In this set of data, criticality is non-numeric factor, which
is converted to a numeric factor (i.e. 1, 0.5 and 0.01 for very critical, moderately critical
and non-critical, respectively) for easier computation.
The membership functions are constructed as per expert’s judgments and
the historical information available for the spares. The trapezoidal membership
function is chosen for all the criteria and Figure 6(a)-(f) shows the diagrammatic
representations of membership functions of input criteria and output parameter.
Table I describes the minimum, maximum and standard deviation values for input
criteria and Table II describes about the membership functions taken for input
and output parameters.
(a) (b) Classification
of spare parts
inventories
467
(c) (d)
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
(e) (f )
Figure 6.
Pictorial
representation of
input and output
Notes: (a) Consumption value; (b) unit price; (c) lead-time; (d) criticality; (e) commonality; parameters
(f ) FRMIC-ABC
Consumption value (Rs) Unit price (Rs) Lead-time (year) Commonality Criticality Table I.
Minimum, maximum
Minimum value 22.22 0.42 1.00 0.01 0.01 and standard
Maximum value 975,323.04 668,192.86 6.00 1.00 1.00 deviation values for
SD 137,532.08 36,597.41 1.44 0.14 0.25 input parameter
All the 243 rules generated are given in Appendix. Figure 7 shows the results of fuzzy
classification. Finally, in Table IV, we have shown the report describing the results of
fuzzy rule based multi-criteria classification and the comparison with traditional ABC
and aggregate scoring model.
Figure 7.
Console showing
the classification
using FRMIC
and FRMIC-C Class (969 items) (see Tables III and IV). Comparing traditional ABC
classification with the proposed FRMIC model, class A items are reduced from 251 to 44
only. But class B spares are increased from 338 to 347 and class C spares are increased
from 771 to 969. It is observed that there is a huge reduction of spares from traditional
A to FRMIC-A classification due to the consideration of multiple criteria. In the new
method, 44 spares require maximum management attention and inspection level, which
is quite beneficial to the organization compared to the cost of managing 251 items in
traditional A class. Similarly, from traditional B class spares (338), FRMIC has shifted
only one spare to FRMIC-A class and 129 spares to FRMIC-B and rest 208 to FRMIC-C
class. Traditional C class spares (771) moved to only FRMIC-B and FRMIC-C classes.
Considering the aggregate scoring model (AGABC), the result shows that AGABC-A has
53 spares, AGABC-B has 686 and AGABC-C has 621 spares. It is observed that counts of
different class spares in aggregate scoring are approximately matching with FRMIC
model as both of them have used multiple criteria for classification. FRMIC gives a better
result as compared to both traditional and aggregate scoring classifications as this model
incorporates the expert’s judgment for classification of spares. AGABC model uses equal
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
Sl. no. Spare code Unit price (Rs) Consumption value (Rs) Lead-time (year) Criticality Commonality TRABC AGABC Defuzzified value FRMIC
Fuzzy-rule-based
of spare parts
Table III.
spare parts
classification for
469
multi-criteria ABC
JQME weights to each of the criteria and ignores any relationships between the criteria whereas
21,4 rule-based approach eliminates these problems. FRMIC methodology can take the decision
from experts in the form of linguistic value of criterion, compute the relationships between
criteria and do the fuzzy-rule-based classifications. Thus, the proposed FRMIC model
gives better output than AGABC and traditional ABC classification.
Here, we have also tried to determine the total cost of spares to each group for different
470 classification methods. As suggested by Diallo et al. (2009), we have used (S 1, S)
inventory policy, also known as base-stock policy for A class spares (s, Q), policy for B
class spares and EOQ policy for C class spares. Table V shows the total cost for each
group of these three classifications. The total cost of 1,360 spares is Rs 102,879,991.35
for the traditional ABC classification, Rs 86,118,720.39 for aggregate scoring model
and Rs 85,020,205.38 for FRMIC model. We observe that total cost for FRMIC is
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
approximately same as that of aggregate scoring model and both these models give
16 percent less cost compared to traditional ABC method. Considering only A class spares
for AGABC and FRMIC, total costs are almost same (Rs 70,757,947.54 and
Rs 68,872,734.50, respectively) and similarly, in the case for B class spares (Rs
14,346,416.36 and Rs 14,833,770.46, respectively). Hence FRMIC model for classification of
spares is valid and adds a new value toward spare parts inventory classifications literature.
Traditional ABC
A (S−1, S) 251 98,478,649.66
B (s, Q) 338 3,726,729.46
C EOQ 771 674,612.23
Optimal total cost of all spares 102,879,991.35
Aggregate scoring
A (S−1, S) 53 70,757,947.54
B (s, Q) 686 14,346,416.36
C EOQ 621 1,014,356.49
Optimal total cost of all spares 86,118,720.39
Table V.
Total cost (Rs) FRMIC
comparison between A (S−1, S) 44 68,872,734.50
proposed model B (s, Q) 347 14,833,770.46
and other C EOQ 969 1,313,700.41
classification models Optimal total cost of all spares 85,020,205.38
easily integrated to any customized ERP applications, which can be accessible any Classification
where in the organization for the decision makers. This model gives the flexibility for of spare parts
inventory management experts to provide their inputs from several places of inventory
stocks scattered in the whole organizations with the help of internet technology. As it is
inventories
stated before, spare parts are the major portion of the total inventory and obsolescence,
commonality, unit price and criticality are crucial factors and considering all these
together, this model is best suited for spare parts inventory classification. The accuracy 471
of the model can be improved gradually by taking the inputs from inventory
management experts at regular interval.
One requires high inspection level, less forecast errors, accurate safety stock levels
and lead-time management approach for FRMIC-based A class items as these items are
highly important spares. The continuous review inventory policy should be applied to
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
References
Bhattacharya, A., Sarkar, B. and Mukherjee, S.K. (2007), “Distance-based consensus method for
ABC analysis”, International Journal of Production Research, Vol. 45 No. 15, pp. 3405-3420.
Braglia, M., Grassi, A. and Montanari, R. (2004), “Multi-attribute classification method for spare
parts inventory management”, Journal of Quality in Maintenance Engineering, Vol. 10
No. 1, pp. 55-65.
Chen, J. (2011), “Peer-estimation for multiple criteria ABC inventory classification”, Computers &
Operations Research, Vol. 38 No. 12, pp. 1784-1791.
Diallo, C., Ait-Kadi, D. and Chelbi, A. (2009), “Integrated spare parts management”, Handbook of
Maintenance Management and Engineering, Springer, Book Part - IV, pp. 191-222.
Flores, B.E. and Whybark, D.C. (1986), “Multiple criteria ABC analysis”, International Journal of
Operations & Production Management, Vol. 6 No. 3, pp. 38-46.
Flores, B.E. and Whybark, D.C. (1987), “Implementing multiple criteria ABC analysis”, Journal of
Operations Management, Vol. 7 Nos 1/2, pp. 79-86.
Flores, B.E., Olson, D.L. and Dorai, V.K. (1992), “Management of multicriteria inventory
classification”, Mathematics and Computer Modeling, Vol. 16 No. 12, pp. 71-82.
Gajpal, P.P., Ganesh, L.S. and Rajendran, C. (1994), “Criticality analysis of spare parts using the
analytic hierarchy process”, International Journal of Production Economics, Vol. 35 Nos 1-3,
pp. 293-297.
Gelders, L.F. and Van Looy, P.M. (1978), “An Inventory policy for slow and fast movers in a
petrochemical plant: a case study”, The Journal of the Operational Research Society, Vol. 29
No. 9, pp. 867-874.
Guvenir, A.H. and Erel, E. (1998), “Multicriteria inventory classification using a genetic
algorithm”, European Journal of Operational Research, Vol. 105 No. 1, pp. 29-37.
JQME Hadi-Vencheha, A. and Mohamadghasemib, A. (2011), “Fuzzy AHP – DEA approach for multiple
criteria ABC inventory classification”, Expert Systems with Applications, Vol. 38 No. 4,
21,4 pp. 3346-3352.
Mamdani, E.H. and Assilian, S. (1975), “An experiment in linguistic synthesis with a fuzzy logic
controller”, International Journal of Man – Machine Studies, Vol. 7 No. 1, pp. 1-13.
Molenaers, A., Baetsc, H., Pintelona, L. and Waeyenbergha, G. (2012), “Criticality classification of
472 spare parts: a case study”, International Journal of Production Economics, Vol. 140 No. 2,
pp. 570-578.
Ng, W.L. (2007), “A simple classifier for multiple criteria ABC analysis”, European Journal of
Operational Research, Vol. 177 No. 1, pp. 344-353.
Ozan, C. and Canbolat, M.S. (2008), “A web-based decision support system for multi-criteria
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
inventory classification using fuzzy AHP methodology”, Expert Systems with Applications,
Vol. 35 No. 3, pp. 1367-1378.
Partovi, F.Y. and Burton, J. (1993), “Using the analytic hierarchy process for ABC analysis”,
International Journal of Operations & Production Management, Bradford, Vol. 13 No. 9,
pp. 29-44.
Ramanathan, R. (2006), “ABC inventory classification with multiple-criteria using weighted linear
optimization”, Computers & Operations Research, Vol. 33 No. 4, pp. 695-700.
Syntetos, A.A., Keyes, M. and Babai, M.Z. (2009), “Demand categorisation in a European spare
parts logistics network”, International Journal of Operations & Production Management,
Vol. 29 No. 3, pp. 292-316.
Tsai, C. and Yeh, S. (2008), “A multiple objective particle swarm optimization approach for
inventory classification”, International Journal of Production Economics, Vol. 114 No. 2,
pp. 656-666.
Wazed, M.A., Ahmed, S. and Yusoff, N. (2009), “Commonality and its measurement in
manufacturing resources planning”, Journal of Applied Sciences, Vol. 9 No. 1, pp. 69-78.
Williams, T.M. (1984), “Stock control with sporadic and slow-moving demand”, The Journal of the
Operational Research Society, Vol. 35 No. 10, pp. 939-948.
Zadeh, L.A. (1965), “Fuzzy sets”, Information and Control, Vol. 8 No. 3, pp. 338-353.
Zadeh, L.A. (1975), “Fuzzy logic and approximate reasoning”, Springer, Vol. 30 No. 3,
pp. 407-428.
Zhou, P. and Fan, L. (2007), “A note on multi-criteria ABC inventory classification using weighted
linear optimization”, European Journal of Operational Research, Vol. 182 No. 3,
pp. 1488-1491.
Further reading
Boylan, J.E., Synteos, A.A. and Karakostas, G.C. (2006), “Classification for forecasting and stock
control: a case study”, Journal of Operational Research Society, Vol. 59, pp. 473-481.
Kuo, R.J., Chi, S.C. and Kao, S.S. (2002), “A decision support system for selecting convenience
store location through integration of fuzzy AHP and artificial neural network”, Computers
in industry, Vol. 47 No. 2, pp. 199-214.
Partovi, F.Y. and Anandarajan, M. (2002), “Classifying inventory using an artificial neural
network approach”, Computers and Industrial Engineering, Vol. 41 No. 4, pp. 389-404.
Silver, E.A., Pyke, D.F. and Peterson, R. (1998), Inventory Management and Production Planning
and Scheduling, Wiley, New York, NY.
Appendix Classification
of spare parts
inventories
Rule no. Unit price (Rs) Consumption value (Rs) Lead-time (year) Criticality Commonality ABC
1 HP HV HL HC HCOM HI
2 HP HV HL HC MCOM HI 473
3 HP HV HL HC LCOM HI
4 HP HV ML HC HCOM HI
5 HP HV ML HC MCOM HI
6 HP HV ML HC LCOM HI
7 HP HV LL HC HCOM HI
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
8 HP HV LL HC MCOM HI
9 HP HV LL HC LCOM MI
10 HP HV HL MC HCOM HI
11 HP HV HL MC MCOM HI
12 HP HV HL MC LCOM HI
13 HP HV ML MC HCOM HI
14 HP HV ML MC MCOM HI
15 HP HV ML MC LCOM MI
16 HP HV LL MC HCOM HI
17 HP HV LL MC MCOM MI
18 HP HV LL MC LCOM MI
19 HP HV HL LC HCOM HI
20 HP HV HL LC MCOM HI
21 HP HV HL LC LCOM MI
22 HP HV ML LC HCOM HI
23 HP HV ML LC MCOM MI
24 HP HV ML LC LCOM MI
25 HP HV LL LC HCOM MI
26 HP HV LL LC MCOM MI
27 HP HV LL LC LCOM LI
28 HP MV HL HC HCOM HI
29 HP MV HL HC MCOM HI
30 HP MV HL HC LCOM HI
31 HP MV ML HC HCOM HI
32 HP MV ML HC MCOM HI
33 HP MV ML HC LCOM MI
34 HP MV LL HC HCOM HI
35 HP MV LL HC MCOM MI
36 HP MV LL HC LCOM MI
37 HP MV HL MC HCOM HI
38 HP MV HL MC MCOM HI
39 HP MV HL MC LCOM MI
40 HP MV ML MC HCOM HI
41 HP MV ML MC MCOM MI
42 HP MV ML MC LCOM MI
43 HP MV LL MC HCOM MI
44 HP MV LL MC MCOM MI
45 HP MV LL MC LCOM LI
46 HP MV HL LC HCOM HI
47 HP MV HL LC MCOM MI
Table AI.
Fuzzy-rule-base
(continued ) (243 rules)
JQME Rule no. Unit price (Rs) Consumption value (Rs) Lead-time (year) Criticality Commonality ABC
21,4
48 HP MV HL LC LCOM MI
49 HP MV ML LC HCOM MI
50 HP MV ML LC MCOM MI
51 HP MV ML LC LCOM LI
52 HP MV LL LC HCOM MI
474 53 HP MV LL LC MCOM LI
54 HP MV LL LC LCOM LI
55 HP LV HL HC HCOM HI
56 HP LV HL HC MCOM HI
57 HP LV HL HC LCOM MI
58 HP LV ML HC HCOM HI
Downloaded by University of Tennessee at Knoxville At 03:12 13 March 2016 (PT)
59 HP LV ML HC MCOM MI
60 HP LV ML HC LCOM MI
61 HP LV LL HC HCOM MI
62 HP LV LL HC MCOM MI
63 HP LV LL HC LCOM LI
64 HP LV HL MC HCOM HI
65 HP LV HL MC MCOM MI
66 HP LV HL MC LCOM MI
67 HP LV ML MC HCOM MI
68 HP LV ML MC MCOM MI
69 HP LV ML MC LCOM LI
70 HP LV LL MC HCOM MI
71 HP LV LL MC MCOM LI
72 HP LV LL MC LCOM LI
73 HP LV HL LC HCOM MI
74 HP LV HL LC MCOM MI
75 HP LV HL LC LCOM LI
76 HP LV ML LC HCOM MI
77 HP LV ML LC MCOM LI
78 HP LV ML LC LCOM LI
79 HP LV LL LC HCOM LI
80 HP LV LL LC MCOM LI
81 HP LV LL LC LCOM LI
82 MP HV HL HC HCOM HI
83 MP HV HL HC MCOM HI
84 MP HV HL HC LCOM HI
85 MP HV ML HC HCOM HI
86 MP HV ML HC MCOM HI
87 MP HV ML HC LCOM MI
88 MP HV LL HC HCOM HI
89 MP HV LL HC MCOM MI
90 MP HV LL HC LCOM MI
91 MP HV HL MC HCOM HI
92 MP HV HL MC MCOM HI
93 MP HV HL MC LCOM MI
94 MP HV ML MC HCOM HI
95 MP HV ML MC MCOM MI
96 MP HV ML MC LCOM MI
97 MP HV LL MC HCOM MI
98 MP HV LL MC MCOM MI
110 MP MV HL HC MCOM HI
111 MP MV HL HC LCOM MI
112 MP MV ML HC HCOM HI
113 MP MV ML HC MCOM MI
114 MP MV ML HC LCOM MI
115 MP MV LL HC HCOM MI
116 MP MV LL HC MCOM MI
117 MP MV LL HC LCOM LI
118 MP MV HL MC HCOM HI
119 MP MV HL MC MCOM MI
120 MP MV HL MC LCOM MI
121 MP MV ML MC HCOM MI
122 MP MV ML MC MCOM MI
123 MP MV ML MC LCOM LI
124 MP MV LL MC HCOM MI
125 MP MV LL MC MCOM LI
126 MP MV LL MC LCOM LI
127 MP MV HL LC HCOM MI
128 MP MV HL LC MCOM MI
129 MP MV HL LC LCOM LI
130 MP MV ML LC HCOM MI
131 MP MV ML LC MCOM LI
132 MP MV ML LC LCOM LI
133 MP MV LL LC HCOM LI
134 MP MV LL LC MCOM LI
135 MP MV LL LC LCOM LI
136 MP LV HL HC HCOM HI
137 MP LV HL HC MCOM MI
138 MP LV HL HC LCOM MI
139 MP LV ML HC HCOM MI
140 MP LV ML HC MCOM MI
141 MP LV ML HC LCOM LI
142 MP LV LL HC HCOM MI
143 MP LV LL HC MCOM LI
144 MP LV LL HC LCOM LI
145 MP LV HL MC HCOM MI
146 MP LV HL MC MCOM MI
147 MP LV HL MC LCOM LI
148 MP LV ML MC HCOM MI
149 MP LV ML MC MCOM LI
161 MP LV LL LC MCOM LI
162 MP LV LL LC LCOM LI
163 LP HV HL HC HCOM HI
164 LP HV HL HC MCOM HI
165 LP HV HL HC LCOM MI
166 LP HV ML HC HCOM HI
167 LP HV ML HC MCOM MI
168 LP HV ML HC LCOM MI
169 LP HV LL HC HCOM MI
170 LP HV LL HC MCOM MI
171 LP HV LL HC LCOM LI
172 LP HV HL MC HCOM HI
173 LP HV HL MC MCOM MI
174 LP HV HL MC LCOM MI
175 LP HV ML MC HCOM MI
176 LP HV ML MC MCOM MI
177 LP HV ML MC LCOM LI
178 LP HV LL MC HCOM MI
179 LP HV LL MC MCOM LI
180 LP HV LL MC LCOM LI
181 LP HV HL LC HCOM MI
182 LP HV HL LC MCOM MI
183 LP HV HL LC LCOM LI
184 LP HV ML LC HCOM MI
185 LP HV ML LC MCOM LI
186 LP HV ML LC LCOM LI
187 LP HV LL LC HCOM LI
188 LP HV LL LC MCOM LI
189 LP HV LL LC LCOM LI
190 LP MV HL HC HCOM HI
191 LP MV HL HC MCOM MI
192 LP MV HL HC LCOM MI
193 LP MV ML HC HCOM MI
194 LP MV ML HC MCOM MI
195 LP MV ML HC LCOM LI
196 LP MV LL HC HCOM MI
197 LP MV LL HC MCOM LI
198 LP MV LL HC LCOM LI
199 LP MV HL MC HCOM MI
200 LP MV HL MC MCOM MI
212 LP MV ML LC MCOM LI
213 LP MV ML LC LCOM LI
214 LP MV LL LC HCOM LI
215 LP MV LL LC MCOM LI
216 LP MV LL LC LCOM LI
217 LP LV HL HC HCOM MI
218 LP LV HL HC MCOM MI
219 LP LV HL HC LCOM LI
220 LP LV ML HC HCOM MI
221 LP LV ML HC MCOM LI
222 LP LV ML HC LCOM LI
223 LP LV LL HC HCOM LI
224 LP LV LL HC MCOM LI
225 LP LV LL HC LCOM LI
226 LP LV HL MC HCOM MI
227 LP LV HL MC MCOM LI
228 LP LV HL MC LCOM LI
229 LP LV ML MC HCOM LI
230 LP LV ML MC MCOM LI
231 LP LV ML MC LCOM LI
232 LP LV LL MC HCOM LI
233 LP LV LL MC MCOM LI
234 LP LV LL MC LCOM LI
235 LP LV HL LC HCOM LI
236 LP LV HL LC MCOM LI
237 LP LV HL LC LCOM LI
238 LP LV ML LC HCOM LI
239 LP LV ML LC MCOM LI
240 LP LV ML LC LCOM LI
241 LP LV LL LC HCOM LI
242 LP LV LL LC MCOM LI
243 LP LV LL LC LCOM LI Table AI.
Corresponding author
Professor S.P. Sarmah can be contacted at: spsarmah@iem.iitkgp.ernet.in
For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com