Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

Advances in Engineering Software 42 (2011) 821–829

Contents lists available at ScienceDirect

Advances in Engineering Software


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

Computer-aided maintenance management systems selection based on a fuzzy AHP


approach
Orlando Durán
Pontificia Universidad Católica de Valparaíso, Av. Los Carrera 01567, Quilpue, Chile

a r t i c l e i n f o a b s t r a c t

Article history: Computerized maintenance management systems (CMMS) are common in today’s industries. CMMS can
Received 5 August 2008 bring a large number of benefits, which include increased productivity, reduced costs, and effective uti-
Received in revised form 7 January 2010 lization of the assets in any manufacturing and service producer. The list of CMMS that are available in
Accepted 13 May 2011
the market has grown very rapidly during the last years. When purchasing a system, one that suits the
Available online 12 June 2011
specific needs and objectives of the company’s maintenance operations should be preferred. Several
selection methods were proposed in literature. Up to now, no article has considered ambiguity and
Keywords:
uncertainty factors when selecting effective CMMS. In addition, CMMS selection decisions involve the
Computerized maintenance management
software
simultaneous consideration of multiple criteria, including tangible and intangible factors; prioritizing
Analytic hierarchy process these factors can be a great challenge and a complex task. Therefore, no attempt has been made to incor-
Fuzzy-AHP porate fuzziness into multicriteria decision-making in the area of CMMS selection. This work proposes a
Software selection fuzzy–based methodology for comparative evaluation of a number of CMMS alternatives. The proposal is
Triangular fuzzy numbers based on the well-known multicriteria decision method called Analytical Hierarchy Process (AHP) with
Comparison of fuzzy numbers triangular numbers. An example is given to illustrate the proposed methodology. Finally, a software pro-
totype for implementing this method was implemented. To illustrate and validate the proposed approach
and the software prototype developed some details are presented and discussed.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction Selection and evaluation of a CMMS is a very difficult and com-


plex task. The following five factors can be identified as the main
The increase in automation and the reduction in inventories in causes of this complexity:
industries have clearly put more pressure on the maintenance sys-
tems. Any disruption to production flows becomes costly and crit- (1) The tremendous number of software products available in
ical. This makes the maintenance function relevant to operations the market.
management to keep organizations productive and profitable (2) The continual advancements and improvements in informa-
along time. Therefore, computerized maintenance management tion technology (IT).
systems (CMMS) are becoming increasingly important in the last (3) The existence of incompatibilities between various hard-
few years. Using CMMS is a highly relevant issue in a production ware and software systems.
environment where the number of critical equipment is high or (4) The functional dissimilarities are difficult to evaluate among
where the need for maintenance resources management is signifi- software packages.
cant. A large variety of computer software is available on the mar- (5) The users lack the technical knowledge and experience for
ket for maintenance management. It is not surprising that many software selection decision making.
companies have been disappointed with the results of their imple-
mented CMMS. An extensive survey [1] reported that there is a As it was previously commented, decision making in the field of
paradox in CMMS selection and implementations. According to this maintenance management software selection has become more
survey, 62% of the respondents changed their maintenance work complex due to a large number of software products in the market,
process to fit the CMMS characteristics and 66% customized the ongoing improvements in information technology, and multiple
CMMS to fit the work process. These numbers reflect that the selec- and sometimes conflicting objectives. Some methodologies and
tion of the most suitable CMMS is a crucial task to eliminate all frameworks for CMMS selection and evaluation have been devel-
these problems and difficulties. oped. Raouf et al. [2] presented an instrument to select suitable
system using a comparative strategy and the concept of relative
E-mail address: orlando.duran@ucv.cl importance among a set of required functions in accordance to

0965-9978/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.advengsoft.2011.05.023
822 O. Durán / Advances in Engineering Software 42 (2011) 821–829

the intended use of the CMMS. More recently, Carnero and Noves work order creation and control, preventive maintenance planning
[3] presented an evaluation system for the selection of computer- and scheduling, human resources, purchasing and receiving, in-
ized maintenance management software in industrial plants using voices matching, and, reporting.
multicriteria methods. Braglia et al. [4] proposed a methodology to Considering Carnero and Novés’ [3] opinion the key functions
perform a selection of the best suited CMMS within process indus- for any CMMS are:
tries. To improve the effectiveness of the methodology proposed by
Braglia et al. [4], they combined AHP with a sensitivity analysis.  Easy work management.
Those evaluation systems use the AHP method in its basic version  Planning function.
(crisp numbers). Up to now, no article has considered ambiguity  Scheduling function.
and uncertainty factors when selecting effective CMMS. In addi-  Budget/cost function.
tion, software selection decisions involve the simultaneous consid-  Spares management.
eration of multiple criteria, including tangible and intangible  Key performance indicators.
factors; prioritizing these factors can be a great challenge and a
complex task. Labib [5] published an investigation of the character- More recently, Bradshaw [18] listed the basic capabilities of a
istics of CMMS, identified their deficiencies. In addition, Labib pro- CMMS. They are: assets database, maintenance activities records,
posed a model to aid the decision analysis capability in CMMS corrective maintenance, preventive maintenance and maintenance
under selection [5]. work scheduling and control. Besides, the same author incorpo-
These factors arise as the main motivation of this research work, rates what he called improved CMMS capabilities; they are inte-
which is aimed at how to select an appropriate CMMS facing the gration and interfacing capabilities and, communication, data
strategic and operational requirements of the organization using collection and transfer.
a multicriteria decision method incorporating concepts of uncer- Wei et al. [19] distinguish two categories of attributes to select
tainty and uncompleted information. an enterprise resource planning system, including system factors
In other words, this study proposes a comprehensive CMMS and vendor factors. Among the subcriteria related to the system
selection framework in which the objective hierarchy is con- factor they suggest the total cost, implementation time, functional-
structed and the appropriate attributes are specified using fuzzy ity, user friendliness, flexibility and reliability. On the other hand,
numbers to provide guidance for CMSS evaluation. The analytic the subcriteria related to the vendor factor they used vendor repu-
hierarchy process (AHP) method [6] and fuzzy numbers are applied tation, technical capability and supplying ongoing service.
for dealing with the ambiguities involved in the assessment of These characteristics lead the authors to propose the hierarchy
CMMS alternatives and relative importance weightings of used in the evaluation approach. Table 1 presents the detailed
attributes. description of the used attributes in the fuzzy AHP model.
Since all organizations are different, it is very important to per-
form a previous identification of maintenance management (MM)
2. Multicriteria decision model IT requirements. Kans [20] presents a formal method for the iden-
tification of MM IT requirements. Accordingly Kans’ work, the
The analytic hierarchy process (AHP) developed by Saaty [6] is a choice of technical features of MMIT is dependent upon the needs
decision-making tool that can handle unstructured or semi struc- and characteristics of the specific organization.
tured decisions with multiperson and multicriteria inputs. It is a Once the IT requirements identification is completed, a series of
decision-rule model that relaxes the measurement of related fac- candidate software systems will arise as a result of the require-
tors to subjective managerial inputs on multiple criteria. AHP has ments identification phase. Those candidates software systems will
several advantages, including its acceptance of inconsistencies in have to be classified or ranked using a MCDM with the participa-
managerial judgments/perceptions and its user friendliness be- tion of domain (maintenance management) experts. For that objec-
cause users may directly input judgment data directly without tive, a fuzzy AHP approach was developed and applied to the
the requirement of mathematical knowledge. It also allows users problem of CMMS selection. The next section discusses the fuz-
to structure complex problems in the form of a hierarchy or a set zy-AHP methodology.
of integrated levels. One of the main advantages of this method
is the relative ease with which it handles multiple criteria. In addi-
tion to this, AHP is easier to understand and it can effectively han- 3. Fuzzy AHP methodology
dle both qualitative and quantitative data. The use of AHP does not
involve cumbersome mathematics. AHP involves the principles of The fuzzy AHP methodology extends Saaty’s AHP by combining
decomposition, pair wise comparisons, and priority vector genera- with fuzzy set theory. In the fuzzy AHP, fuzzy ratio scales are used
tion and synthesis. The power of AHP has been validated by empir- to indicate the relative strength of the factors in the corresponding
ical application in diverse areas such as healthcare [7], planning criteria. Therefore, a fuzzy judgment matrix can be constructed.
[8], mining [9], project management [10], missile systems [11], The final scores of alternatives are also represented by fuzzy num-
new product development [12] and manufacturing [13]. In addi- bers. The optimum alternative is obtained by ranking the fuzzy
tion AHP has been used in making decisions that involve ranking, numbers using special algebra operators.
selection, evaluation, and selection of machines and IT based sys- The next three steps can summarize the procedure of applying
tems [14–16]. fuzzy AHP:
To construct the hierarchy of objectives and attributes of a
generic CMMS an extended review of the related literature was i. Construction of a hierarchical structure for the problem to be
conducted. This review focused on CMMS selection and, on gen- solved.
eral-purpose software selection problem. Two works arise as the ii. Establish the fuzzy judgment matrix and a fuzzy weight
most significant ones to the goal pursuit by our approach. vector.
Cato and Mobley [17] listed some activities which constitute iii. Rank all alternatives and select the optimal one.
subsystems or modules of a generic CMMS: equipment/asset re-
cords creation and maintenance, asset bills of materials creation Three levels compose the hierarchy of the evaluation system.
and maintenance, asset and work order history, inventory control, The suggested hierarchy is depicted in Fig. 1. The first level is the
O. Durán / Advances in Engineering Software 42 (2011) 821–829 823

Table 1 numbers to indicate the relative contribution or impact of each


Attribute details. alternative on a criterion, a fuzzy judgment vector is then obtained
Factor Attributes Sub attributes for each criterion. The fuzzy judgment matrix A is built with all the
Functionality  Module completion fuzzy judgment vectors. The weight vector W is used to represent
 Function fitness the decision maker’s opinion of the relative importance of each cri-
 Maintenability terion during the decision process.
System Flexibility  Upgrade ability Though the purpose of AHP is to capture the expert’s knowl-
 Ease of integration
 Ease in-house customization
edge, the conventional AHP still cannot reflect the human thinking
Friendliness  Ease operation style. In spite of its popularity, this method is often criticized be-
 Ease learning cause of a series of pitfalls associated with the AHP technique. They
 Low customization can be summarized as follows:
Implementation  Implementation time
Difficulties  Implementation costs
Reputation and Stability  Scale of vendor  Its inability to adequately handle the inherent uncertainty and
 Financial condition imprecision associated with the mapping of the decision-
 Market share (and internationali- maker’s perception to exact numbers [8].
zation)  In the traditional formulation of the AHP, human’s judgments
Vendor Good technical capability  R&D capability
are represented as exact (or crisp, according to the fuzzy logic
 On line support
 Experience in the industry terminology) numbers. However, in many practical cases the
Service  Warranties human preference model is uncertain and decision-makers
 Consultant services might be reluctant or unable to assign exact numerical values
 Training service
to the comparison judgments.
 Proximity
 Although the use of the discrete scale of 1–9 has the advantage
of simplicity, the AHP does not take into account the uncer-
tainty associated with the mapping of one’s judgment to a
goal of the problem, that is, select the most suitable computer number.
based maintenance management system. The two main decision
criteria are placed in the second level. They are system and vendors In order to overcome the aforementioned shortcomings, a fuzzy
factors. In the third level are the most relevant sub-criteria for each extension of AHP, was developed to solve the hierarchical fuzzy
one of the criteria listed in level 2. problems. In the next sections a fuzzy AHP technique is proposed,
In this methodology, triangular fuzzy numbers represents all and an example for the evaluation and justification of advanced
elements in the judgment matrix and weight vectors. Using fuzzy manufacturing system is presented.

Functionality

Flexibility

System
Factors

Friendliness
CMMS 1

CMMS 2

Selection of the most Implementation


suitable CMMS CMMS 3
Reputation

CMMS n
Technic. Cap.

Vendor
Factors

Service

Fig. 1. Hierarchy of the CMMS selection problem.


824 O. Durán / Advances in Engineering Software 42 (2011) 821–829

A fuzzy number ~ x expresses the meaning ‘about x0 . Each mem- subjective pairwise comparisons or capture certain degree of vague-
bership function is defined by three parameters of the symmetric ness (Table 2).
triangular fuzzy number, (l, m, r), the left point, middle point and We know that matrix A is a real and positive matrix. As well as,
right point of the range over which the function is defined. Fuzzy since aij = 1/aji, if i is not equal to j, A is a reciprocal matrix.
membership function and the definition of a fuzzy number are Next, the eigenvector, eigenvalue and the IC index are calcu-
shown in Fig. 2. lated, now taking these parameters as fuzzy numbers.
8 To estimate the fuzzy eigenvector from A matrix the next equa-
> 1 x¼m
>
> tion is used:
< xl l6x6m
ml
lðxÞ ¼ ð1Þ !1=n
>
>
nx
m6x6n Y
n
>
: nm Vi ¼ ~ij
a ð6Þ
0 otherwise j¼1

When the decision-maker faces a complex and uncertain prob-


Therefore, we have
lem and expresses his/her comparison judgments as uncertain ra-
tios, such as ‘about two times more important’, ‘between two and ~11  a
V 1 ¼ ða ~1n Þ1=n
~12  a13      a ð7Þ
four times less important’, etc. the standard AHP steps, and spe-
cially, eigenvalue prioritization approach, cannot be considered
~n1  a
V n ¼ ða ~n2  a ~nn Þ1=n
~n3      a ð8Þ
as straightforward procedures. Indeed, the assessment of local pri-
orities, based on pair wise comparisons needs some prioritization
Eigenvector V is compound by the n triangular numbers defined as:
method to be applied. Next a brief description about addition, mul-
tiplication and division of triangular numbers is given. The fuzzy V ¼ ðV1; V2; . . . VnÞ
operators were adapted from [21] and are based on the extent
analysis with the use of triangular fuzzy numbers for pairwise where Vi is a triangular number defined as (Vl, Vm, Vu)
comparison scale [22]. As the traditional AHP methodology, eigenvector is to be nor-
Let A and B be two triangular fuzzy numbers, with their param- malized according the next relation:
eters as follows: T ¼ ðw1=Rwi w2=Rwi w3=Rwi ... wn=RwiÞ ð9Þ
e ¼ ða1; a2; a3Þ
A That is, by dividing each element of the preference matrix with
the sum of its respective column each element of the normalized
e ¼ ðb1; b2; b3Þ
B eigenvector can be obtained.
Where T is the normalized eigenvector. From this normalized
Then, fuzzy numbers multiplication is defined by:
eigenvector the priorities or importance of the attributes under
eB
A e ¼ ða1  b1; a2  b2; a3  b3Þ ð2Þ analysis is extracted.
In order to control the result of the method, the consistency ra-
At the other hand, fuzzy numbers division is defined as follows: tio need to be calculated. The deviations from consistency are ex-
e Be ¼ ða1=b3; a2=b2; a3=b1Þ pressed by the following equation:
A= ð3Þ
kmax  n
whilst the reciprocal value of a triangular fuzzy number (a, b, c) is CI ¼ ð10Þ
given by (1/a, 1/b, 1/c). The power of a triangular fuzzy number is
n1
given by The consistency ratio (CR) is used to estimate directly the con-
sistency of pairwise comparisons. The CR is computed by dividing
e n ¼ ða1; a2; a3Þn ¼ ða1n ; b2n ; c3n Þ
A ð4Þ the CI by a value obtained from the table of Random Consistency
As can be seen in Fig. 3, the relative importance of a number Index (RI) created by Saaty
over other fuzzy number is gradual and not abrupt. CR ¼ CI=RI ð11Þ
~ i be a set of decision maker’s opinion of the relative impor-
Let w
tance of the one alternative over other one. The meaning of each If the CR is less than 10%, the comparisons are acceptable, other-
fuzzy number is defined in Table 2. wise not. RI represents the average index for randomly generated
Using this scale we have the comparison matrix A, e where aij ele- weights.
ments represent the estimative of the wi/wj relation. Since kmax is a triangular number, it has to be defuzzified into a
crisp number to compute the CI. We suggest here using the central
~ w1
w1= ~ w2=
~ w1
~ . . . wn= ~
~ w1 value of kmax, because of the symmetry of the triangular number,
e ¼ w1=
A ~ w2
~ w2=
~ w2
~ . . . wn= ~
~ w2 ð5Þ the central value corresponds to the centroid of the triangular area.
~ ~
~ w2=wn
~ . . . wn=
~ wn
~ As an alternative, Leung and Cao [23] propose a fuzzy consistency
w1=wn
definition with consideration of a tolerance deviation. Essentially,
Experts’ judgments or preferences among the options using Saatýs the fuzzy ratios of relative importance, allowing certain tolerance
scale is represented now by triangular numbers to express deviation, are formulated as constraints on the membership values

~ ~ ~ ~ ~
1 3 5 7 9

Fig. 2. Membership function of a triangular number. Fig. 3. Saatýs scale expressed as fuzzy sets.
O. Durán / Advances in Engineering Software 42 (2011) 821–829 825

Table 2 Considering this method and Eq. (9), the second element of the
Saaty’s scale expressed in fuzzy numbers. eigenvector is the highest value and corresponds to the third attri-
Relative importance Definition bute (AT3) operation easiness.
~ Equal importance The normalization process yields a new form of the eigenvector
1
~
3 Weak importance in which each entry is a triangular number, as follows:
~
5 Strong importance
T ¼ ðð0:02 0:11 1:09Þð0:02 0:16 1:39Þð0:04 0:37 2:82Þ
~
7 Demonstrated importance over the other
~
9 Absolute importance  ð0:03 0:29 2:20Þð0:01 0:05 0:36Þð0:004 0:02 0:17ÞÞ
For testing the consistency of the resulting eigenvector, Saaty
proposed the following relation:

of the local priorities. The fuzzy local and global weights are deter- kmax ¼ T  w ð13Þ
mined via the extension principle. where w is computed by the sum of the columns of the preferences
matrix .

4. Case study W ¼ ðð5:3 11:3 19:5Þð4:5 8:7 15:6Þð1:8 2:4 5:9Þð2:6 3:9 8:9Þ
 ð16:1 24:2 34:3Þð25:0 35:0 45:0ÞÞ
In this section, in order to prove the applicability and validity
the proposed methodology is applied to a case study. An invest- Next kmax is calculated by
ment decision in CMSS technology of a given manufacturer was ta- kmax ¼ ð0:55 6:55 98:94Þ
ken into consideration. Suppose that a set of three experts provide
the following fuzzy preference relations on a set of three alterna- Then, to calculate the CI (crisp) we used the central value of the
tives CMMS (CMMS1, CMMS2, CMMS3). This case study supposes triangular number kmax.
that functional aspects are covered by the three alternatives of CI ¼ ð6:55  6Þ=5 ¼ 0:11
CMMS. They are in equal conditions regarding aspects as module
completion, function fitness, maintainability and other functional In addition, CR is computed
aspects. This is corroborated in [24] that reports that main func- CR ¼ 0:11=1:24 ¼ 0:089 < 0:10
tionality elements all belong to the core of a CMMS and are almost
always available regardless if the user asks for them or not. There- This proves total consistency of the evaluations expressed by
fore, other intangible aspects are the factors that were considered the comparisons matrix.
by the methodology. Based on the weight vector (eigenvector) the priorities or rela-
As it is known, it may be impractical to make paired compari- tive importance of the attributes were obtained by ranking the
sons among CMMSs with respect to every detailed dimension or eigenvector values. The ranked order of the six attributes is as fol-
sub attribute of the hierarchy. The difficulty arises because too lows: AT3, AT4, AT2, AT1, AT5 and AT6. Next, the three possible
many attributes lead to numerous paired comparisons in AHP CMMS were compared with respect each of the six attributes.
and may cause an inefficient process. Therefore, a simplified model The corresponding fuzzy pairwise comparison matrices are shown
was formulated. After a set of interviews, a series of six qualitative in Table 4.
attributes was selected to perform the analysis. The six attributes Next, we can find the scores of the alternative CMMS with re-
are: flexibility, operation easiness, reliability, quality, implementa- spect to the six attributes, which are shown in Table 5.
tion easiness and maintainability (Fig. 4). This six attributes are The local weights of all CMMS for each attribute are obtained by
represented by the six following symbols: AT1, AT2, AT3, AT4, AT5 multiplying their relative weights by the weights of the attributes.
and AT6 respectively. Once the decisions makers performed the Table 6 shows these local weights.
pair wise comparisons for the set of attributes A matrix is obtained The overall classification can be obtained by multiplying (trian-
(Table 3). This comparison matrix is constructed by using Saaty’s gular product) the weights matrix (Table 6) by the transposed
scale but now with triangular numbers. eigenvector of the attributes (Table 5). Table 7 shows the overall
To find the relative importance or priorities of the six attributes classification vector.
eigenvector, eigenvalue and the RC index are to be computed. Thus, Thus the priority scores for the CMMS alternatives are obtained,
the eigenvector (with triangular values) is as follows: and they are ranked based on their magnitude (using Eq. (12)).

V ¼ ðð0:12 0:3 1:08Þð0:15 0:42 1:38Þð0:35 1:00 2:80Þ CMMS1 1.43


CMMS2 1.73
 ð0:27 0:77 2:18Þð0:05 0:12 0:35Þð0:03 0:06 0:17ÞÞ
CMMS3 2.60

Before to proceed to perform the normalizations is necessary an


additional fuzzy ranking procedure in order to compare fuzzy
scores and to obtain a linear order among them. There is a number Thus, CMMS 3 must be selected by the users or recommended
of procedures to perform the ranking process [23] and more re- by the AHP fuzzy methodology.
cently [25], among them we propose, because of its simplicity,
the utilization of the representative method, which is given by 5. Proposed software
the following relation:
As can be easily proved, AHP with fuzzy numbers requires many
b ¼ a1 þ 2a2 þ a3
A ð12Þ
time-consuming calculations. Depending upon the number of attri-
4 butes and alternatives taken into consideration a lot of time is nec-
essary to make all calculations in order to reach the final solution.
where A = (a1, a2, a3) is a triangular number and b
a represents the As the number of attributes increases, the dimension of the prob-
representative ordinal of a triangular number. lem expands. This could lead to a great number of mathematical
826 O. Durán / Advances in Engineering Software 42 (2011) 821–829

Service

Technic. Cap. CMMS 1

Flexibility
Selection of the most
suitable CMMS CMMS 2

Reputation

Friendliness CMMS 3

Implementation

Fig. 4. Abbreviated hierarchy.

Table 3
Comparisons matrix of the attributes considered for selection of CMMS.

AT1 AT2 AT3 AT4 AT5 AT6


AT1 (1, 1, 3) (1/5, 1/3, 1) (1/6, 1/4, 1/2) (1/5, 1/3, 1) (1, 3, 5) (5, 7, 9)
AT2 (1, 3, 5) (1, 1, 3) (1/5, 1/3, 1) (1/6, 1/4, 1/2) (1, 3, 5) (7, 9, 9)
AT3 (2, 4, 6) (1, 3, 5) (1, 1, 3) (1/7, 1/5, 1/3) (1/7, 1/5, 1/3) (1, 3, 5)
AT4 (1, 3, 5) (2, 4, 6) (3, 5, 7) (1, 1, 3) (1/5, 1/3, 1) (3, 5, 7)
AT5 (1/5, 1/3, 1) (1/5, 1/3, 1) (3, 5, 7) (1, 3, 5) (1, 1, 3) (3, 5, 7)
AT6 (1/9, 1/7, 1/5) (1/9, 1/9, 1/7) (1/5, 1/3, 1) (1/7, 1/5, 1/3) (1/7, 1/5, 1/3) (1, 1, 3)

Table 4
Fuzzy pairwise comparisons for the alternative CMMSs.

AT1 AT2
CMMS1 CMMS2 CMMS3 CMMS1 CMMS2 CMMS3
CMMS1 (1.0 1.0 3.0) (0.2 0.33 1.0) (0.2 0.33 1.0) CMMS1 (1.0 1.0 3.0) (0.33 1.0 1.0) (0.33 1.0 1.0)
CMMS2 (1.0 3.0 5.0) (1.0 1.0 3.0) (1.0 3.0 5.0) CMMS2 (1.0 1.0 3.0) (1.0 1.0 3.0) (0.33 1.0 1.0)
CMMS3 (1.0 3.0 5.0) (0.2 0.33 1.0) (1.0 1.0 3.0) CMMS3 (1.0 1.0 3.0) (1.0 1.0 3.0) (1.0 1.0 3.0)
AT3 AT4
CMMS1 (1.0 1.0 3.0) (0.2 0.33 1.0) (0.11 0.11 0.14) CMMS1 (1.0 1.0 3.0) (7.0 9.0 9.0) (0.11 0.11 0.14)
CMMS2 (1.0 3.0 5.0) (1.0 1.0 3.0) (0.11 0.14 0.2) CMMS2 (0.11 0.11 0.14) (1.0 1.0 3.0) (0.14 0.2 0.33)
CMMS3 (7.0 9.0 9.0) (5.0 7.0 9.0) (1.0 1.0 3.0) CMMS3 (7.0 9.0 9.0) (3.0 5.0 7.0) (1.0 1.0 3.0)
AT5 AT6
CMMS1 (1.0 1.0 3.0) (0.2 0.33 1.0) (0.33 1.0 1.0) CMMS1 (1.0 1.0 3.0) (4.0 6.0 8.0) (0.33 1.0 1.0)
CMMS2 (1.0 3.0 5.0) (1.0 1.0 3.0) (1.0 3.0 5.0) CMMS2 (1.0 3.0 5.0) (1.0 1.0 3.0) (0.33 1.0 1.0)
CMMS3 (1.0 1.0 3.0) (0.2 0.33 1.0) (1.0 1.0 3.0) CMMS3 (1.0 1.0 3.0) (1.0 1.0 3.0) (1.0 1.0 3.0)

Table 5
Eigenvectors of the CMMS alternatives with respect to the six attributes.

VAT1 ((0.06 0.22 1.05) (0.10 0.46 1.79) (0.08 0.32 1.37)) Table 7
VAT2 ((0.09 0.33 0.82) (0.11 0.33 0.99) (0.13 0.33 1.19)) Overall classification vector (with triangular numbers).
VAT3 ((0.08 0.17 0.40) (0.11 0.25 0.55) (0.29 0.58 1.14))
CMMS1 (0.01 0.25 5.20)
VAT4 ((0.15 0.29 0.58) (0.08 0.16 0.33) (0.26 0.55 1.11))
CMMS2 (0.01 0.27 6.38)
VAT5 ((0.07 0.27 1.03) (0.10 0.46 1.76) (0.08 0.27 1.23))
CMMS3 (0.03 0.48 9.39)
VAT6 ((0.13 0.38 0.95) (0.10 0.34 0.88) (0.12 0.28 0.97))

Table 6
Local weights for the CMMS alternatives with respect to the six attributes.

AT1 AT2 AT3 AT4 AT5 AT6


CMMS1 (0.06 0.22 1.05) (0.09 0.33 0.82) (0.08 0.17 0.40) (0.15 0.29 0.58) (0.07 0.27 1.03) (0.13 0.38 0.95)
CMMS2 (0.10 0.46 1.79) (0.11 0.33 0.99) (0.11 0.25 0.55) (0.08 0.16 0.33) (0.10 0.46 1.76) (0.10 0.34 0.88)
CMMS3 0.08 0.32 1.37 (0.13 0.33 1.19) (0.29 0.58 1.14) (0.26 0.55 1.11) (0.08 0.27 1.23) (0.12 0.28 0.97)
O. Durán / Advances in Engineering Software 42 (2011) 821–829 827

Fig. 5. Screenshot of preference matrix.

Fig. 6. First phase analysis results.

and fuzzy operations. Therefore, software aid may be very useful to the comparison matrix is entirely filled with importance values
automatically carry out the fuzzy AHP process. A software proto- (using fuzzy scale and slide bars) and total consistency is proved,
type for fuzzy AHP application was developed. The software was the system provides the ranking of the attributes according to the
programmed by using MATLAB 7.0 on a PC platform. The operation information input by the user. The system provides the eigenvector
sequence will be demonstrated in the following paragraphs, and the eigenvalue, plus the consistency value (Fig. 6). The system
through the use of several screenshots. presents the eigenvector using crisp numbers. Conversions of fuzzy
Initially, user must select and input the criteria chosen for eval- numbers into crisp numbers are carried out by the software.
uation the CMSS alternatives. The software prototype keeps in a In the second part of the software, the user must input the pair-
database a series of attributes that the user can select to perform wise comparisons between two specific CMMS. This task is made
the comparison analysis. Additionally, the attribute database con- accordingly to each one of the considered attributes (Fig. 7).
tains a set of generic attributes labeled as ‘‘attribute i’’ where ‘‘i’’ Next, the software finds the scores of the alternative CMMSs
stands for the number of a given attribute. Next, the user(s) must with respect to the six attributes. The software displays the results
fill the pairwise comparisons matrix for the attributes. Fig. 5 shows using crisp representation. Therefore, defuzzification of the values
the dialog box where the analyst can input the pairwise compari- is performed and the priority scores for the CMMS alternatives are
sons among software’s attributes. obtained. The ranked list indicates that in this case the CMMS with
To find the relative importance or priorities of the six attributes the higher value has to be selected by the users. In the final screen
eigenvector, eigenvalue and the RC index are to be computed. Once of the system, the results of analysis are shown (Fig. 8).
828 O. Durán / Advances in Engineering Software 42 (2011) 821–829

Fig. 7. Comparison matrix among CMMSs with respect to the attribute Vendor’s market share.

Fig. 8. Final results.

6. Conclusions was found to function satisfactorily. We believe that this method-


ology and the developed software tool is an alternative to both the
In this paper a fuzzy-AHP-based methodology for selecting conventional AHP method as well as other fuzzy-based approaches
Computerized Maintenance Management Software was proposed. for software selection, mainly because of its simplicity and the pos-
In order to take into account the uncertainty and in order to im- sibility of incorporating subjective parameters.
prove imprecision in ranking attributes and/or software alterna-
tives, the presented approach introduces triangular numbers into
traditional AHP method. Adoption of fuzzy numbers allows deci- References
sions-makers to have more freedom of estimation regarding the
overall importance of attributes and real alternatives. The pro- [1] O’Hanlon T. CMMS best practices. Mainten J 2004;17(3):19–22.
posed methodology was tested on a real-world example and was [2] Raouf A, Zulfigar A, Duffuaa SO. Evaluating a computerised maintenance
management system. Int J Operat Prod Manage 1993;13(3):38–48.
found that it functions satisfactorily. We believe that this method- [3] Carnero MC, Novés JL. Selection of computerised maintenance management
ology is a feasible alternative to both the conventional AHP meth- system by means of multicriteria methods. Prod Plann Control
od, as well as, other fuzzy-based approaches for CMMS selection, 2006;17(4):335–54.
[4] Braglia M, Carmignani G, Frosolini M, Grassi A. AHP-based evaluation of CMMS
mainly because of its simplicity and the possibility of incorporating
software. J Manuf Technol Manage 2006;17(5):585–602.
subjective parameters and linguistic terms in expressing main soft- [5] Labib A. A decision analysis model for maintenance policy selection using a
ware characteristics. Additionally, a fuzzy AHP based Software for CMMS. J Qual Mainten Eng (JQME) 2004;10/3:191–202.
selecting computerized maintenance management was proposed. [6] Saaty TL. Analytic hierarchy process. New York: McGraw Hill; 1980.
[7] Liberatore MJ, Nydick RL. The analytic hierarchy process in medical and health
The program was written in MATLAB and run on a desktop PC pow- care decision making: a literature review. Eur J Operat Res
ered by Microsoft Windows XP. It was tested on several tests and 2003;189(1):194–207.
O. Durán / Advances in Engineering Software 42 (2011) 821–829 829

[8] Vahadilla OS, Kumar S. Analytic hierarchy process: an overview of applications. [17] Cato R, Mobley K. Computer-managed maintenance systems in process plants:
Eur J Operat Res 2006;169(1):1–29. a step-by-step guide to effective management of maintenance, labor, and
[9] Kazakidis VN, Mayer Z, Scoble MJ. Decision making using the analytic inventory, 1st ed. Butterworth-Heinemann College; 1998.
hierarchy process in mining engineering. Trans Inst Min Metall Sect A – Min [18] Bradshaw L. Improved CMMS and asset management systems – but do they
Technol 2004;113(1):A30–42. lead to success? Mainten Asset Manage 2005;20(2):21–8.
[10] Mustafa MA, Albahar JF. Project risk assessment using the analytic hierarchy [19] Wei Chun-Chin, Chien Chen-Fu, Wang Mao-Jiun. An AHP-based approach to
process. IEEE Trans Eng Manage 1991;38(1):46–52. ERP system selection. Int J Prod Econ 2005;96:47–62.
[11] Cheng Ching-Hsue. Evaluating naval tactical missile systems by fuzzy AHP [20] Kans M. An Approach for determining the requirements of computerized
based on the grade value of membership function. Eur J Operat Res maintenance management systems. Comput Ind 2008;59:32–40.
1997;96(2):343–50. [21] Chiu C-Y, Park CS. Capital budgeting decisions with fuzzy projects. Eng
[12] Ayag Z. A fuzzy AHP-based simulation approach to concept evaluation in a Economist 1998;43(2):125–50.
NPD environment. IIE Transactions; September 2005. p. 827–42(16). [22] Chang DY. Applications of the extent analysis method on fuzzy AHP. Eur J
[13] Karsak EE, Tolga E. Fuzzy multicriteria decision making procedure for Operat Res 1996;95:649–55.
evaluating advanced manufacturing system in investments. Int J Prod Econ [23] Leung LC, Cao D. On consistency and ranking of alternatives in fuzzy AHP. Eur J
2001;69:49–64. Operat Res 2000;124:102–13.
[14] Bozdag CE, Kahraman C, Ruan D. Fuzzy group decision making for selection [24] Ingwald A, Kans M. The use of IT within maintenance management for
among computer integrated manufacturing systems. Comput Ind continuous improvement. In: Proceedings of MIMAR 2007 – 6th IMA
2003;51:13–29. international conference on modeling in industrial maintenance and
[15] Shamsuzzaman M, Sharif Ullah AMM, Bohez ELJ. Applying linguistic criteria in reliability. p. 51–6.
FMS selection: fuzzy set AHP approach. Integrated Manuf 2003;14/3:247–54. [25] Ciptomulyono U. Fuzzy goal programming approach for deriving priority
[16] Ordoobadi Sharon M, Mulvaney NJ. Development of a justification tool for weights in the analytical hierarchy process (AHP) method. J Appl Sci Res
advanced manufacturing technologies: system-wide benefits value analysis. J 2008;4(2):171–7.
Eng Technol Manage 2001;18(2):157–84.

You might also like