0 views

Original Title: Une en 60300 3-3-2009

Uploaded by orlandoduran

- Seven Plus or Minus Two2
- Solution of Linear System Theory and Design 3ed for Chi-Tsong Chen
- kotak_220711_01
- circulant matrices
- Laurio Ps123longtest2
- diff_eq
- 2008-JVA-GSVD.pdf
- 14180-20150623-ETABS-Main Build
- Jonathan Baier-Getting Started With Kubernetes-Packt Publishing (2017)
- The-Use-Quasi-comparison-Functions-with-Rayleigh-Ritz-Method-Beam.pdf
- Test 1 Linear Algebra
- valuation
- 1893839-AQA-MFP4-QP-JUN12.pdf
- International Journal of Heat and Mass Transfer Study on Steady-state Thermal Condu
- Determining Leaf-Angle Distribution of Vineyard Leaves Using Terrestrial LiDAR
- Linear Systems
- Op Vibration
- Chandana_Anand_scholarly_paper
- Initiation of Coverage
- Math_222-F14

You are on page 1of 8

com

Expert Systems

with Applications

Expert Systems with Applications 34 (2008) 1787–1794

www.elsevier.com/locate/eswa

on a Fuzzy-AHP approach

Orlando Durán *, José Aguilo

Escuela de Ingenierı́a Mecánica, Universidad Católica de Valparaı́so, Av. Los carrera 01567 quilpué, Chile

Abstract

Investment evaluation methods play an important role in today’s competitive manufacturing environment. Shrinking proﬁt margins

and diversiﬁcation require careful analysis of investments and decisions regarding these investments are crucial for the survival of the

manufacturing industry. Both economic evaluation criterion and strategic criteria such as ﬂexibility, quality improvement, which are

not quantitative in nature, are considered for evaluation. Much has been written about the deﬁciencies of traditional models for justifying

advanced manufacturing systems. The use of fuzzy set theory allows incorporating unquantiﬁable, incomplete and partially known infor-

mation into the decision model. In this paper, an analytic hierarchical process (AHP) based on fuzzy numbers multi-attribute method is

proposed for the evaluation and justiﬁcation of an advanced manufacturing system. Finally, an example of machine tool selection is used

to illustrate and validate the proposed approach.

2007 Elsevier Ltd. All rights reserved.

ignore time value of money.

Advanced manufacturing technologies (AMT) is an The conventional DCF methods do not appear to be

important item in the design of a manufacturing system. suitable on their own for the evaluation of AMT invest-

Using proper manufacturing technology can enhance ments due to the non-monetary impacts posed by the man-

the production process, provide eﬀective utilization of ufacturing system. The inadequacy of traditional ﬁnancial

resources, increase productivity and improve system ﬂexibil- justiﬁcation measures lies on their deterministic nature.

ity, repeatability and reliability. Therefore, given the wide The probabilistic cash ﬂow analysis can be used if the

range of advanced manufacturing technologies available probabilities distributions of the possible outcomes are

today, the determination of the best equipment available known. However, when the frequency distribution of the

for a given production scenario is not an easy task. possible outcomes is not known, as in revenues and

Economic justiﬁcation methods of manufacturing invest- expenses of a new production system, most decision-makers

ments have been discussed thoroughly in the past years. are reluctant to employ DCF methods during the eval-

Economic analysis methods are the basic discounted cash uation phase. Table 1 shows an updated version of the

ﬂow (DCF) techniques such as present worth, annual classiﬁcation of the justiﬁcation methods for advanced

worth, internal rate of return, etc., and other techniques manufacturing technologies.

Much has been written about the deﬁciencies of tradi-

tional engineering economic models for justifying AMT

(Bozdag, Kahraman, & Ruan, 2003; Ordoobadi & Mulva-

*

Corresponding author.

ney, 2001; Shamsuzzaman, Sharif Ullah, & Bohez, 2003).

E-mail addresses: orlando.duran@ucv.cl (O. Durán), jmacnat@hot- Most of the dissatisfaction revolves around the following

mail.com (J. Aguilo). points:

0957-4174/$ - see front matter 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.eswa.2007.01.046

1788 O. Durán, J. Aguilo / Expert Systems with Applications 34 (2008) 1787–1794

Table 1

Justiﬁcation methods for advanced manufacturing technologies

Techniques Advantages Disadvantages

Economic • Payback • Ease in data collection • Do not take into account strategic and non-economic beneﬁts

method • Intuitive appeal • Consider a single objective of cash ﬂows, and ignore other beneﬁts

• ROI such as quality and ﬂexibility

• NPV, IRR

Strategic • Technical • Require less technical data • Necessity to use these techniques with economic or analytic ones

importance • Use the general objectives of the ﬁrm since they consider only long-term intangible beneﬁts

• Business

objectives

• Competitive

advantages

• R&D

Analytic • AHP • Uncertainty of the future and the multi- • Require more data

• Mathematical objectivity can be incorporated • Usually more complex than the economic analysis

Programming • Subjective criteria can be introduced in the

• Stochastic modeling phase

methods

• Fuzzy sets

theory

• Short term returns are emphasized rather than long term tiﬁcation is observed in approximately 60% of the manu-

strategy. facturing ﬁrms responding the questionnaire.

• A variety of assumptions are made designed to deal with In the justiﬁcation process of advanced information or

the uncertainty involved in predicting the future manufacturing systems, quantiﬁcation of some of the

environment. revenue or quality improvements is often diﬃcult if not

• And, the range of beneﬁts considered is diminished due impossible. Many managers have argued that accounting

to the diﬃculty in quantifying many important factors. methodologies restrict the adoption and use of advanced

technologies and are incapable of quantifying many of the

According to Shamsuzzaman et al. (2000), numerous beneﬁts oﬀered by these systems in many organizations.

tangible and intangible beneﬁts are expected after the intro- The analytic hierarchy process (AHP) developed by

duction of AMT. Among the intangible beneﬁts one can Saaty (1980) is a decision-making tool that can handle

mention reduced labor costs, high capital utilization, faster unstructured or semistructured decisions with multiperson

throughput times, better quality control, increased safety and multicriteria inputs. It is a decision-rule model that

and a better response to unpredictable situations. relaxes the measurement of related factors to subjective

Sullivan and William (1986) points out the inadequacy managerial inputs on multiple criteria. AHP has several

of traditional ﬁnancial justiﬁcation measures of project advantages, including its acceptance of inconsistencies in

worth such as return on investment, payback, and net pres- managerial judgments/perceptions and its user friendliness

ent worth in considering the strategic merits of advanced because users may directly input judgment data without

manufacturing technologies. further requiring mathematical knowledge. It also allows

According Karsak and Tolga (2001), DCF methods users to structure complex problems in the form of a hier-

appear as the most popular economic justiﬁcation method- archy or a set of integrated levels. AHP can also be com-

ology; however, determining cash ﬂows (revenues, bined with well-known operations research techniques to

expenses) and discount rates as crisp values can lead to handle more diﬃcult problems. One of the main advanta-

erroneous results in most real-life applications. ges of this method is the relative ease with which it handles

The results of the surveys conducted by Leﬂey (1994) for multiple criteria. In addition to this, AHP is easier to

justiﬁcation of advanced manufacturing technology understand and can eﬀectively handle both qualitative

(AMT) in the UK, and by Leﬂey and Sarkis (1997) for and quantitative data. The use of AHP does not involve

appraisal of AMT investments in the UK and US both cumbersome mathematics. AHP involves the principles of

support the diﬃculty in assessing AMT investments due decomposition, pair wise comparisons, and priority vector

to their non-quantiﬁable beneﬁts. As a result of this diﬃ- generation and synthesis. The power of AHP has been val-

culty, over 80% of the respondents in the US and UK point idated by empirical application in diverse areas such as

out that not all potential beneﬁts of AMT investments healthcare, politics, and urban planning (Karsak & Tolga,

are considered in the ﬁnancial justiﬁcation process. 2001). It has been used in making decisions that involve

Furthermore, the results of the surveys state that subjective ranking, selection, evaluation, optimization, and prediction

assessment of AMT investment with/without ﬁnancial jus- (Lee, Lau, Liu, & Tam, 2001).

O. Durán, J. Aguilo / Expert Systems with Applications 34 (2008) 1787–1794 1789

Though the purpose of AHP is to capture the expert’s tive importance of each criterion during the decision

knowledge, the conventional AHP still cannot reﬂect the process.

human thinking style. In spite of its popularity, this A fuzzy number ~x expresses the meaning ‘about x’. Each

method is often criticized because of a series of pitfalls membership function is deﬁned by three parameters of the

associated with the AHP technique which can be summa- symmetric triangular fuzzy number, (l, m, r), left, middle

rized as follows: and right points of the range over which the function is

deﬁned. Fuzzy membership function and the deﬁnition of

• Its inability to adequately handle the inherent uncer- a fuzzy number are shown in Fig. 1.

tainty and imprecision associated with the mapping of When the decision-maker faces a complex and uncertain

the decision-maker’s perception to exact numbers (Lef- problem and expresses his/her comparison judgments as

ley & Sarkis, 1997). uncertain ratios, such as ‘about two times more important’,

• In the traditional formulation of the AHP, human’s ‘between two and four times less important’, etc., the stan-

judgments are represented as exact (or crisp, according dard AHP steps, and specially, eigenvalue prioritization

to the fuzzy logic terminology) numbers. However, in approach, cannot be considered as straightforward proce-

many practical cases the human preference model is dures. Indeed, the assessment of local priorities, based on

uncertain and decision-makers might be reluctant or pair wise comparisons needs some prioritization method

unable to assign exact numerical values to the compari- to be applied. Next a brief description about addition, mul-

son judgments. tiplication and division of triangular numbers is given. The

• Although the use of the discrete scale of 1–9 has the fuzzy operators were adapted from Chiu and Park (1994).

advantage of simplicity, the AHP does not take into Let A and B be two triangular fuzzy numbers, with their

account the uncertainty associated with the mapping parameters as follows:

of one’s judgment to a number.

e ¼ ða1 ; a2 ; a3 Þ

A

In order to overcome the aforementioned shortcomings, e ¼ ðb1 ; b2 ; b3 Þ

B

a fuzzy extension of AHP, was developed to solve the hier-

archical fuzzy problems. In the next sections a Fuzzy-AHP

Then, fuzzy numbers multiplication is deﬁned by

technique is proposed, and an example for the evaluation

and justiﬁcation of advanced manufacturing system is eB

A e ¼ ða1 b1 ; a2 b2 ; a3 b3 Þ ð1Þ

presented.

At the other hand, fuzzy numbers division is deﬁned as

2. Fuzzy-AHP methodology follows:

The Fuzzy-AHP methodology extends Saaty’s AHP by e B

A= e ¼ ða1 =b3 ; a2 =b2 ; a3 =b1 Þ ð2Þ

combining it with the fuzzy set theory. In the Fuzzy-

AHP, fuzzy ratio scales are used to indicate the relative Whilst the reciprocal value of a triangular fuzzy number

strength of the factors in the corresponding criteria. There- (a, b, c) is given by (1/a, 1/b, 1/c). The power of a triangular

fore, a fuzzy judgment matrix can be constructed. The ﬁnal fuzzy number is given by

scores of alternatives are also represented by fuzzy num-

bers. The optimum alternative is obtained by ranking the e n ¼ ða1 ; a2 ; a3 Þn ¼ ðan ; bn ; cn Þ

A ð3Þ

1 2 3

fuzzy numbers using special algebra operators.

The next three steps can summarize the procedure of As seen in Fig. 2, the relative importance of a number over

applying Fuzzy-AHP: other fuzzy number is gradual and not abrupt.

Let w

~ i be a set of decision maker’s opinion of the relative

(i) Construct a hierarchical structure for the problem to importance of the one alternative over other one. The

be solved. meaning of each fuzzy number is deﬁned in Table 2.

(ii) Establish the fuzzy judgment matrix and a fuzzy Using this scale we have the comparison matrix A, e

weight vector. where aij elements represent the estimative of the wi/wj

(iii) Rank all alternatives and select the optimal one. relation

matrix and weight vectors are represented by triangular 1 x=m 1

fuzzy numbers. Using fuzzy numbers to indicate the rela- x−l l≤x≤m (1)

tive contribution or impact of each alternative on a crite- m−l μ(x)

μ(x)=

rion, a fuzzy judgement vector is then obtained for each n−x

n−m m≤x≤n

criterion. The fuzzy judgment matrix A is built with all 0 otherwise

0 l m n

the fuzzy judgement vectors. The weight vector W is used

to represent the decision maker’s opinion of the rela- Fig. 1. Membership function of a triangular number.

1790 O. Durán, J. Aguilo / Expert Systems with Applications 34 (2008) 1787–1794

~

1

~

3

~

5

~

7

~

9 As the traditional AHP methodology, eigenvector is to

be normalized according the next relation

T ¼ ðw1 =Rwi w2 =Rwi w3 =Rwi ; . . . ; wn =Rwi Þ

where T is the normalized eigenvector. From this normal-

Fig. 2. Saaty’s scale expressed as fuzzy sets. ized eigenvector the priorities or importance of the attri-

butes under analysis is extracted.

Table 2 In order to control the result of the method, the consis-

Saaty’s scale expressed in fuzzy numbers tency ratio [need] * needs * to be calculated. The deviations

Relative importance Deﬁnition from consistency are expressed by the following equation:

~

1 Equal importance k max n

~

3 Weak importance CI ¼ ð8Þ

~

n1

5 Strong importance

~

7 Demonstrated importance over the other The consistency ratio (CR) is used to directly estimate the

~

9 Absolute importance consistency of pairwise comparisons. The CR is computed

by dividing the CI by a value obtained from the table of

Random Consistency Index (RI) created by Saaty

f1 =f

w w1 f2 =f

w w1 ... fn =f

w w1

f1 =f

w w2 f2 =f

w w2 ... fn =f

w w2 ð4Þ CR ¼ CI=RI

e¼

A f1 =f

w wn f2 =f

w wn ... fn =f

w wn If the CR is less than 10%, the comparisons are acceptable,

otherwise not. RI represents the average index for ran-

Experts’ judgments or preferences among the options using domly generated weights.

Saaty’s scale is represented now by triangular numbers to Since kmax is a triangular number, it has to be defuzziﬁed

express subjective pairwise comparisons or capture certain into a crisp number to compute the CI. We suggest here

degree of vagueness (Table 3). using the central value of kmax, because of the symmetry

We know that matrix A is a real and positive matrix. As of the triangular number, the central value corresponds

well as, since aij = 1/aji, if i is not equal to j, A is a recipro- to the centroid of the triangular area.

cal matrix.

Next, the eigenvector, eigenvalue and the IC index are 3. Case study

calculated, now taking these parameters as fuzzy numbers.

To estimate the fuzzy eigenvector from A matrix the In this section, the proposed methodology is applied to a

next equation is used case study, in order to prove its applicability and validity.

!1=n

Yn A new CNC turning center investment decision of a given

Vi ¼ aij ð5Þ manufacturer was taken into consideration. A triplet of

j¼1 decisions makers was asked to evaluate a set of three alter-

Therefore, we have natives machine tools (MT1, MT2, MT3). After a set of

interviews, a series of six qualitative attributes was selected

1=n

V 1 ¼ ða

11 a12 a13 a1n Þ ð6Þ to perform the analysis. The six attributes are: ﬂexibility,

operation easiness, reliability, quality, implementation eas-

1=n iness and maintainability. This six attributes are repre-

V n ¼ ða

n1 an2 an3 ann Þ ð7Þ sented by the six following symbols: AT1, AT2, AT3,

Eigenvector Vi is compound by the n triangular numbers AT4, AT5 and AT6 respectively. Once the decision makers

deﬁned as perform the pairwise comparisons for the set of attributes

the A matrix is obtained (Table 3). This comparison matrix

V ¼ ðV 1 ; V 2 ; . . . ; V n Þ

is constructed by using Saaty’s scale but now with triangu-

where Vi is a triangular number deﬁned as (Vl, Vm, Vu). lar numbers.

Table 3

Comparisons matrix of the attributes considered for selection of machine tools

AT1 AT2 AT3 AT4 AT5 AT6

AT1 (1,1,3) (1/5,1,1/3) (5,7,9) (1,3,5) (1,3,5) (5,7,9)

AT2 (1,3,5) (1,1,3) (7,9,9) (1,3,5) (1,3,5) (7,9,9)

AT3 (1/9,1/7,1/5) (1/9, 1/9,1/7) (1,1,3) (1/7, 1/5, 1/3) (1/7, 1/5, 1/3) (1,3,5)

AT4 (1/5,1,1/3) (1/5,1,1/3) (3,5,7) (1,1,3) (1/5,1,1/3) (3,5,7)

AT5 (1/5,1,1/3) (1/5,1,1/3) (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) (1/5,1,1/3) (1/7, 1/5,1/3) (1/7, 1/5,1/3) (1,1,3)

O. Durán, J. Aguilo / Expert Systems with Applications 34 (2008) 1787–1794 1791

To ﬁnd the relative importance or priorities of the six (Buckley, 1985; Mikhailov & Tsvetinov, 2004), among

attributes eigenvector, eigenvalue and the RC index are them we propose the utilization of the representative meth-

to be computed. Thus, the eigenvector (with triangular val- od, which is given by the following relation:

ues) is as follows:

b ¼ a1 þ 2a2 þ a3

A ð9Þ

V ¼ ðð1:3 2:3 4:3Þð1:9 3:6 5:6Þð0:3 0:4 0:6Þð0:6 0:98 2:3Þ 4

where A = (a1,a2,a3) is a triangular number and A b repre-

ð0:8 1:4 3:0Þð0:2 0:2 0:5ÞÞ

sents the representative ordinal of a triangular number.

Considering this method, the second element of the

Before performing the normalizations an additional fuzzy eigenvector is the highest value and corresponds to the sec-

ranking procedure is needed in order to compare fuzzy ond attribute (AT2) operation easiness.

scores and to obtain a linear order among them. There The normalization process yields a new form of the

are a number of procedures to perform the ranking process eigenvector in which each entry is a triangular number,

as follows:

Table 4

Fuzzy pairwise comparisons for the alternative machines

V ¼ ðð0:03 0:26 2:42Þð0:04 0:40 3:17Þð0:01 0:04 0:34Þ

Maq1 Maq2 Maq3 ð0:01 0:11 1:30Þð0:02 0:16 1:70Þð0:004 0:03 0:26ÞÞ

AT1

Maq1 (1.00 1.00 3.00) (0.20 0.33 1.00 ) (0.20 0.33 1.00) For testing the consistency of the resulting eigenvector,

Maq2 (1.00 3.00 5.00) (0.20 0.33 1.00) (1.00 3.00 5.00) Saaty proposed the following relation:

Maq3 (1.00 3.00 5.00) (0.20 0.33 1.00) (0.20 0.33 1.00)

kmax ¼ V w ð10Þ

AT2

Maq1 (1.00 1.00 3.00) (0.33 1.00 1.00) (0.33 1.00 1.00) where w is computed by the sum of the columns of the pref-

Maq2 (1.00 1.00 3.00) (1.00 1.00 3.00) (0.33 1.00 1.00) erences matrix

Maq3 (1.00 1.00 3.00) (1.00 1.00 3.00) (1.00 1.00 3.00)

W ¼ ðð2:6 5:0 10:4Þð1:8 2:2 6:3Þð19:2 27:3 36:0Þ

AT3

Maq1 (1.00 1.00 3.00) (3.00 5.00 7.00) (1.00 1.00 3.00) ð4:3 10:4 18:7Þð3:5 7:7 14:7Þð20:0 30:0 40:0ÞÞ

Maq2 (0.14 0.20 0.33) (1.00 1.00 3.00) (0.14 0.20 0.33)

Maq3 (0.33 1.00 1.00) (3.00 5.00 7.00) (1.00 1.00 3.00) Next kmax is calculated by

AT4 kmax ¼ ð4:13 6:47 12:73Þ

Maq1 (1.00 1.00 3.00) (7.00 9.00 9.00) (5.00 7.00 9.00)

Maq2 (0.11 0.11 0.14) (1.00 1.00 3.00) (0.14 0.20 0.33) Then, to calculate the CI (crisp) we used the central value

Maq3 (0.11 0.14 0.20) (3.00 5.00 7.00) (1.00 1.00 3.00) of the triangular number kmax.

AT5 CI ¼ ð6:47 6Þ=5 ¼ 0:09

Maq1 (1.00 1.00 3.00) (0.20 0.33 1.00) (0.33 1.00 1.00)

Maq2 (1.00 3.00 5.00) (1.00 1.00 3.00) (1.00 3.00 5.00) In addition, RC is computed

Maq3 (1.00 1.00 3.00) (0.20 0.33 1.00) (1.00 1.00 3.00)

RC ¼ 0:09=1:24 ¼ 0:07 < 0:10

AT6

Maq1 (1.00 1.00 3.00) (4.00 6.00 8.00) (2.00 4.00 6.00) This proves the total consistency of the evaluations ex-

Maq2 (1.00 3.00 5.00) (1.00 1.00 3.00) (0.20 0.30 1.00) pressed by the comparisons matrix. Based on the weight

Maq3 (0.17 0.25 0.50) (1.00 3.00 5.00) (1.00 1.00 3.00) vector (eigenvector) the priorities or relative importance

of the attributes are as follows: AT2, AT1, AT5, AT4,

AT3 and AT6. Next, the three possible turning centers

were compared with each of the six attributes. The corre-

Table 5

sponding fuzzy pairwise comparison matrices are shown

Eigenvectors of the machine alternatives and their six attributes

in Table 4.

VAT1 ((0.1 0.2 1.0) (0.1 0.5 1.8) (0.1 0.3 1.4))

Next, we can ﬁnd the scores of the alternative machines

VAT2 ((0.1 0.3 0.8) (0.1 0.3 1.0) (0.1 0.3 1.2))

VAT3 ((0.2 0.4 1.2) (0.1 0.2 0.5) (0.1 0.4 1.0)) and their six attributes, shown in Table 5.

VAT4 ((0.3 0.6 1.1) (0.10.2 0.3) (0.1 0.3 0.6)) The local weights of all machines for each attribute are

VAT5 ((0.1 0.3 1.0) (0.1 0.5 1.8) (0.1 0.3 1.2)) obtained by multiplying their relative weights by the

VAT6 ((0.2 0.5 1.3) (0.1 0.3 0.9) (0.1 0.3 0.8)) weights of the attributes. Table 6 shows these local weights.

Table 6

Local weights for the machine alternatives and their six attributes

AT1 AT2 AT3 AT4 AT5 AT6

Maq1 (0.06 0.22 1.05) (0.09 0.33 0.82) (0.16 0.41 1.22) (0.29 0.57 1.10) (0.07 0.27 1.03) (0.17 0.47 1.27)

Maq2 (0.10 0.46 1.79) (0.11 0.33 0.99) (0.07 0.18 0.51) (0.08 0.15 0.32) (0.10 0.46 1.76) (0.09 0.27 0.87)

Maq3 0.08 0.32 1.37 0.13 0.33 1.19 (0.13 0.41 1.01) (0.13 0.27 0.56) (0.08 0.27 1.23) (0.09 0.26 0.78)

1792 O. Durán, J. Aguilo / Expert Systems with Applications 34 (2008) 1787–1794

Overall classiﬁcation vector (with triangular numbers) based on the sequence of the steps shown before.

Maq1 (0.01 0.33 9.07) Initially, the user must input the selected evaluation cri-

Maq2 (0.01 0.36 11.28) teria into the machine tool alternatives. The software pro-

Maq3 (0.01 0.31 10.46)

totype keeps a database with a series of about twelve

attributes that the user can select to perform the compari-

The overall classiﬁcation can be obtained by multiplying son analysis. Additionally, the attribute database contains

(triangular product) the weights matrix (Table 6) by the a set of eight generic attributes labeled as ‘‘attribute I’’

transposed eigenvector of the attributes (Table 6). Table where I stands for the number of a given attribute. Next,

7 shows the overall classiﬁcation vector. the user(s) must ﬁll in the pairwise comparisons matrix

Thus the priority scores for the machine alternatives are for the attributes. Secondly, the software solves and

obtained, and they are ranked based on their magnitude obtains an/the eigenvector, eigenvalue and CI leading to

the priority weight for each attribute. The second part of

Maq1 2:43 the software deals with the identiﬁcation of the machine

Maq2 3:00 ð11Þ tool alternative that best suits the attributes selected by

Maq3 2:77 the user(s), considering the priority weights obtained

for the attribute level (part I of the software). The results

Thus, Machine 3 must be selected by the users or recom- of the analysis process are presented to the user in a speciﬁc

mended by the AHP fuzzy methodology. screen as shown in Fig. 3. In conventional AHP the pair-

wise comparisons are made using crisp numbers and ratios.

The developed software allows the user to express the

4. Proposed software degree of uncertainty associated with the mapping of one’s

perception or judgement to a fuzzy scale using sliders in

As can be easily seen, AHP with fuzzy numbers requires each comparisons matrix cell.

many time-consuming calculations. Depending on the Once the comparison matrix is entirely ﬁlled with impor-

number of attributes and alternatives taken into consider- tance values (using fuzzy scale), the system provides the

ation, a lot of time is necessary to make all calculations eigenvector and eigenvalue, plus the consistency value

in order to reach the ﬁnal solution. As the number of attri- (Fig. 4).

butes increases, the dimension of the problem expands. If total consistency is proved, the system provides the

This could lead to a great number of mathematical and ranking of the attributes according to the information

fuzzy operations. Therefore, software aid may be very use- input by the user. In the second part of the software, the

ful to automatically carry out the Fuzzy-AHP process. user must input the pair wise comparisons between two

A software prototype for Fuzzy-AHP application was speciﬁc machine tools. This task is made according each

developed. The software was programmed using MAT- of the considered attributes (Fig. 5).

O. Durán, J. Aguilo / Expert Systems with Applications 34 (2008) 1787–1794 1793

Fig. 5. Comparison matrix among projects for each one of the criteria. Fig. 6. Final results.

In the ﬁnal screen of the system, the results of analysis method and as well as other Fuzzy based approaches for

are shown in Fig. 6. machine tool selection, mainly because of its simplicity

and the possibility of incorporating subjective parameters.

5. Conclusions

References

In this paper a Fuzzy-AHP based Software for selecting

Bozdag, C. E., Kahraman, C., & Ruan, D. (2003). Fuzzy group decision

machine tools was proposed. In order to consider uncer- on making for selection among computer integrated manufacturing

tainty and improving imprecision in ranking attributes systems. Computers in Industry, 51(2003), 13–29.

and/or machine alternatives. The presented approach Buckley, J. J. (1985). Ranking alternatives using fuzzy numbers. Fuzzy

introduces triangular numbers into traditional AHP Sets and Systems, 15(1), 21–31.

Chiu, C. Y., & Park, C. S. (1994). Fuzzy cash ﬂow analysis using present

method. Adoption of fuzzy numbers allows decisions mak-

worth criterion. The Engeneering Economist, 39(2), 113–137.

ers to achieve a better estimation ﬂexibility regarding the Karsak, E. E., & Tolga, E. (2001). Fuzzy multicriteria decisión making

overall importance of attributes and real alternatives. For procedure for evaluating advanced manufacturing system in invest-

the methodology of Fuzzy-AHP explained above a pro- ments. International Journal of Production economics, 69(2001), 49–64.

gram was written in MATLAB and run on a desktop PC Lee, W. B., Lau, H., Liu, Z., & Tam, S. (2001). A fuzzy analytic hierarchy

process approach in modular product design. Expert Systems, Febru-

powered by Microsoft Windows XP. It was tested on the

ary, 18(1), 32–42.

above example and other situations and was found to work Leﬂey, F. (1994). Capital investment appraisal of advanced manufacturing

properly. We believe that this methodology and software technology. International Journal of Production Research, 32(12),

tool is a viable alternative to both the conventional AHP 2751–2776.

1794 O. Durán, J. Aguilo / Expert Systems with Applications 34 (2008) 1787–1794

Leﬂey, F., & Sarkis, J. (1997). Short-termism and the appraisal of AMT Shamsuzzaman, M. (2000). Selection of A FMS based on fuzzy set theory

capital projects in the US and UK. International Journal of Production and AHP methods. ISE-Thesis, Asian Institute of technology,

Research, 35(2), 341–368. Bangkok.

Mikhailov, L., & Tsvetinov, P. (2004). Evalution of services using a fuzzy Shamsuzzaman, M., Sharif Ullah, A. M. M., & Bohez, E. L. J. (2003).

analytic herarchy process. Applied Soft Computing, 5(2004), 23–33. Applying linguistic crieteria in FMS selection: fuzzy set AHP

Ordoobadi, S. M., & Mulvaney, N. J. (2001). Development of a approach. Integrated manufacturing (14/3), 247–254.

justiﬁcation tool for advanced manufacturing technologies: system- Sullivan & William, G. (1986). Justifying new technology: models ies can

wide beneﬁts value analysis. Journal of Engineering and Technology use to include strategic, non-monetary factors in automation decisions.

Management, 18(2), 157–184. Industrial Engineering, 18(3), 42–50.

Saaty, T. L. (1980). Analytic Herarchy Process. New York: McGraw Hill.

- Seven Plus or Minus Two2Uploaded byfauzisd
- Solution of Linear System Theory and Design 3ed for Chi-Tsong ChenUploaded byadambose1990
- kotak_220711_01Uploaded byBinay Kumar
- circulant matricesUploaded byAbeer Zahid
- Laurio Ps123longtest2Uploaded byJohn Kelly Mercado
- diff_eqUploaded byHan Seo
- 2008-JVA-GSVD.pdfUploaded byIgor Gjorgjiev
- 14180-20150623-ETABS-Main BuildUploaded byPurushotam Tapariya
- Jonathan Baier-Getting Started With Kubernetes-Packt Publishing (2017)Uploaded byTarun Mittal
- The-Use-Quasi-comparison-Functions-with-Rayleigh-Ritz-Method-Beam.pdfUploaded byGabriel Ferrario
- Test 1 Linear AlgebraUploaded byReginaldRemo
- valuationUploaded byDeep Arora
- 1893839-AQA-MFP4-QP-JUN12.pdfUploaded byastargroup
- International Journal of Heat and Mass Transfer Study on Steady-state Thermal ConduUploaded byla zeki
- Determining Leaf-Angle Distribution of Vineyard Leaves Using Terrestrial LiDARUploaded byjpeterson1
- Linear SystemsUploaded byChernet Tuge
- Op VibrationUploaded byMing Hsiang Cheng
- Chandana_Anand_scholarly_paperUploaded bygarumat10
- Initiation of CoverageUploaded byluxfranca
- Math_222-F14Uploaded byPamela Phelps
- LongFin - A Global FinTech HouseUploaded byLongFin Corp
- ps0Uploaded byAni
- 12.pdfUploaded byali
- Wind Buckling of Metal Tanks During Their ConstructionUploaded byARNOUX
- SFD5_AvanziniUploaded bydurak89
- euler_notesUploaded byRavi Duggirala
- 188437221 Case 22 Victoria Chemicals Plc AUploaded byriders29
- 2007 December JP Morgan MA Reference ManualUploaded byRohitGuleria
- Final Exam Study GuideUploaded byChris Lin
- Eigenvalues and Eigenvectors - MATLAB EigUploaded byAkash Ramann

- norma española.pdfUploaded byRichard Catacora
- TESIS DANIELA Contreras.pdfUploaded byorlandoduran
- 1-s2.0-S0957417411013509-main.pdfUploaded byorlandoduran
- quiz 7 procesamiento de elastomeros.docxUploaded byorlandoduran
- Quiz 10 de Procesos III Sistemas de manufactura .docxUploaded byorlandoduran
- paperUploaded byorlandoduran
- ayudantia eval proyUploaded byorlandoduran
- Informe Memoria Cristobal Tobar.docxUploaded byorlandoduran
- Procesos de FabricacionUploaded byorlandoduran
- Preventivo Neumaticos TTCUploaded byorlandoduran
- cursocnc.pdfUploaded byOtrebligRabocse
- Newsletter Diplomado Gestioìn de la ingenieriìa de Mantenimiento DGIM-final.pdfUploaded byorlandoduran
- Clases Parte I .pdfUploaded byorlandoduran
- sustainability-10-00362(4).pdfUploaded byorlandoduran
- DIAGRAMA DE FLUJO Clasificación de las llantas de una flota de vehículos de TPS.pdfUploaded byorlandoduran
- Simulador LambdaUploaded byorlandoduran
- Ejemplo Distribucion NormalUploaded byorlandoduran
- mrp-ho1Uploaded byorlandoduran
- 3-2-1_acabados_rugosidadUploaded byDiego Gutiérrez Martínez
- Preparación y Presentación de Proyectos de InversiónUploaded byapi-3748966
- Tablas de EquivalenciaUploaded byorlandoduran
- APUNTES Evaluacion de Proyectos 2_2Uploaded byorlandoduran
- 2016 Advances in Tool Wear in Turning ProcessUploaded byorlandoduran
- Sustainability 10 02260Uploaded byorlandoduran
- Ejemplos Planes de Procesos Mario RossiUploaded byorlandoduran
- 139 Haanstra and Braaksma Life Cycle Costing in Physical Asset Management a Multiple Case StudyUploaded byorlandoduran
- 20180412123454526Uploaded byorlandoduran
- EXAMEN Procesos Fabricacion I 2017Uploaded byorlandoduran
- pronosticos ejemplos resueltosUploaded byorlandoduran

- Gardening RobotUploaded byRoyalAryans
- mvc-130330091359-phpapp01Uploaded byAdv Sunil Joshi
- Konkreet Labs - How-To OSC-MIDI for Windows (and Mac) UsersUploaded byYungaro David Santalla Aquim
- Online Payment _ Passport SevaUploaded byAbdul Hakeem Semar Kamaluddin
- Dos 5 MemoryUploaded byrnb_zounds
- Agile Model and PrototypingUploaded bynaspuloy
- 10-01004Uploaded byrachmat99
- csa-xe-3-13s-asr-920-book1G-10G-ports (1).pdfUploaded bySyed Jailany
- PptUploaded bymiteshmvjce
- AspenPolymers+Vol1V7_3-UsrUploaded byFirman Suryalaga Dikusuma
- 05 01 Ra41205en16agla0 Rrm OverviewUploaded byPham Duy Nhat
- Face Detection in Images- Neural Networks & Support Vector MachinesUploaded bySuleman Jamil
- How to Write Unmaintainable CodeUploaded bytan_73
- Md-100 Exam QuestionsUploaded byalex5187
- Konfiguracija Terminal SERVERAUploaded bymbnaming
- Critical Issues in SAP End to END ImplementationUploaded bySandeshBelkunde
- PigUploaded byvrkatevarapu
- LightningUploaded byproteus555
- Dell PowerEdge M520 server solution: Energy efficiency and database performanceUploaded byPrincipled Technologies
- One Way ANOVA.pdfUploaded bySarah Seunarine
- CDS – One Concept, Two Flavors _ SAP BlogsUploaded byManoj Mahajan
- Cehv9 BrochureUploaded byAyahnyaAzkaAzmi
- E1926Uploaded bygrats_singco
- 8051 Student ProjectsUploaded byAbhijay Sisodia
- How to Scribus - CardsUploaded byKerry Moore
- DocuNet Viewer IOSUploaded byIlluga Ta Phrom
- Project Oberon, part 1, SystemUploaded byPablo Oscar Cayuela
- Automatic Energy Meter Billing System with Theft DetectionUploaded byIRJET Journal
- (1) Jack Huizenga's Answer to How Do You Prove That a Number is a Transcendental NumberUploaded byKrutarth Patel
- Faxpress PDFUploaded byAmber