Você está na página 1de 22

Applied Soft Computing 10 (2010) 10401061

Contents lists available at ScienceDirect

Applied Soft Computing


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

A new approach to mining method selection based on modifying the


Nicholas technique
Amir Azadeh a,*, M. Osanloo b, M. Ataei c
a
Department of Mining Engineering, Engineering Faculty of Azad University, Sciences & Researches Branch, PO Box 14515/775, Tehran, Iran
b
Department of Mining, Metallurgy and Petroleum Engineering, Amirkabir University of Technology, Hafez Ave., PO Box 15875-4413, Tehran, Iran
c
Department of Mining, Geophysics and Petroleum Engineering, Shahrood University of Technology, 7th tir Sq., PO Box 36155-316, Shahrood, Iran

A R T I C L E I N F O A B S T R A C T

Article history: The main purpose of this paper is to represent a solution to the problem of mining method selection
Received 16 January 2009 (MMS) in mining projects. To this aim, the well-known MMS technique of Nicholas has been modied so
Received in revised form 8 July 2009 that in addition to eliminate its defects, it would be possible for mining engineers to easily assign their
Accepted 5 September 2009
engineering judgments to unsteady and uncertain characteristics of mineral resources. So, in order to
Available online 25 September 2009
resolve the problems of weighting of the Nicholas technique, analytic hierarchy process (AHP) as the
most similar multi-criteria decision making (MCDM) tool to Nicholas technique was applied. Due to
Keywords:
inability of crisp numbers for assigning of decision maker (DM) judgments to ambiguities of mineral
Mining method selection
resources, trapezoidal fuzzy numbers also were used for better modeling of those ambiguities. Moreover,
AHP
Fuzzy sets theory a two-step algorithm containing hierarchical technicaloperational model (HTOM) and also hierarchical
Nicholas technique economical model (HEM), inspired by Nicholas technique, was proposed. These models include some
Optimum prot new criteria which are added to the Nicholas technique. Therefore using fuzzy AHP (FAHP), mining
alternatives are rstly ranked based on HTOM and then, the most protable of those alternatives is
selected by the HEM. As a case study, the north anomaly of Choghart iron mine was used to compare the
proposed approach with the Nicholas technique. The results indicated that the proposed approach
eliminated the problems of Nicholas technique. Proposed approach also introduces a protable mining
alternative to start the mining operations. It should be applied to avoid further feasibility studies in
mining projects.
2009 Elsevier B.V. All rights reserved.

1. Introduction extraction of that resource is started by the selected method, it is


not possible to change the method and replace it with another one.
Mining methods are those operating methods that are used in Because such a replacement is usually so costly that the whole
order to extract mineral resources from the earth. Considering the project could become uneconomical. MMS is therefore an
complications of the geometrical and geological characteristics of irreversible stage in mine planning. On the other hand, selecting
mineral resources, no single mining method can be used for the a mining method for mineral resources is completely dependent on
extraction of all the mineral resources. So, taking the unique the uncertain geometrical and geological characteristics of the
characteristics of each mineral resource into account it is necessary resource. This uncertainty is such that no certain value could be
to use just the suitable mining method for the extraction of a assigned for none of these characteristics. For example a certain
certain resource, so that the applied method would have the slope or volume may not be assigned to mineral resources using
maximum technicaloperational congruence with the geometrical crisp numbers. The above-mentioned issues indicate the impor-
and geological conditions of the mineral resource. It is also tance and complication of MMS in mining projects. The sensitivity
obligatory for the applied mining method to be protable in of MMS in mining projects has led to different solutions introduced
comparison with the other methods [44]. The process of selecting a by different researchers. Lack of a systematic solution in 70 and 80
mining method for the extraction of mineral resources is called decades caused the introduction of some qualitative solutions
MMS. After MMS process is done for a mineral resource and the introduced by Boshkov et al. [9], Morrison [41], Laubscher [33],
Hamrin [25], Brady and Brown [10], Hartman [26], Adler and
Thompson [1], to nd a solution for MMS problem. In these studies,
MMS procedure has been looked from qualitative viewpoint. This
* Corresponding author at: No. 7, Western 19th Street, Modares Blvd., Taavon,
Semnan City, Semnan, Iran. Tel.: +98 2314441582. means that MMS procedure is done by linguistic and qualitative
E-mail address: amirazadeh@gmail.com (A. Azadeh). denitions. Furthermore, selection criteria which are used in these

1568-4946/$ see front matter 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.asoc.2009.09.002
A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061 1041

studies are insufcient. As another problem of them, these


methodologies can be often applied for specic category of mining
methods only, such as underground mining methods. They had
used some owchart, tabular denitions and graphs as methodol-
ogy to select a mining method or a suitable set of them. Generally,
based on these classier methods, some mining methods can be
participated into selection process. These methodologies were not
adequate to solve the MMS problems.
In 1981, the rst numerical approach to MMS was represented
Fig. 1. Trapezoidal fuzzy number A a=b; g =d [11]. by Nicholas. Using some criteria and some mining methods and

Table 1
Fundamental scales to pair-wise comparison in AHP [4,49].

Numerical intensity Linguistic denitions Descriptions


of importance

1 Equally preferred Two activities contribute equally to the objective


2 Equally to moderately Can be used when compromise is needed between 1 and 3
3 Moderately preferred Experience and judgment strongly favor one activity over another
4 Moderately to strongly Can be used when compromise is needed between 3 and 5
5 Strongly preferred Experience and judgment strongly favor one activity over another
6 Strongly to very strongly Can be used when compromise is needed between 5 and 7
7 Very strongly preferred An activity is strongly favored, its dominance demonstrated in practice
8 Very strongly to extremely Can be used when compromise is needed between 7 and 9
9 Extremely preferred The evidence favoring one activity over another is of tile highest
possible order of afrmation

Table 2
Average random consistency (RI) [49].

Size of matrix 1 2 3 4 5 6 7 8 9 10

Random consistency 0 0 0.58 0.9 1.12 1.24 1.32 1.41 1.45 1.49

Table 3
Weighting procedure of the Nicholas MMS technique: ore geometry attributes [43,44].

Alternatives Criteria

General shape Ore thickness Ore plunge Grade distribution


a a a a a a a a a a
M T/P I N I T VT F I S Ua Ga Ea

Open pit mining 3 2 3 2 3 4 4 3 3 4 3 3 3


Block caving 4 2 0 49 0 2 4 3 2 4 4 2 0
Sublevel stoping 2 2 1 1 2 4 3 2 1 4 3 3 1
Sublevel caving 3 4 1 49 0 4 4 1 1 4 4 2 0
Long wall mining 49 4 49 4 0 49 49 4 0 49 4 2 0
Room and pillar 0 4 2 4 2 49 49 4 1 0 3 3 3
Shrinkage stoping 2 2 1 1 2 4 3 2 1 4 3 2 1
Cut and ll 0 4 2 4 4 0 0 0 3 4 3 3 3
Top slicing 3 3 0 49 0 3 4 4 1 2 4 2 0
Stull stoping 0 2 4 4 4 1 1 2 3 3 3 3 3

M: massive; T/P: tabular or platy; I: irregular; N: narrow (<10 m); I: intermediate (<1030 m); T: thick (<30100 m); VT: very thick (<100 m); F: at (<208); I: intermediate
(20558); S: steep (>558); U: uniform; G: gradational; E: erratic.
a
Mining methods.

Fig. 2. Fuzzy numbers based on trapezoidal membership of A a=b; g =d which use in fuzzy hierarchy analysis: (a) triangular, (b) more than a to 1, (c) less than d to 1, (d)
between a/1 and g/1, (e) at least a/1, and (f) at most d/1 [11,12].
1042 A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061

Table 4
Weighting procedure of the Nicholas MMS technique: Ore zone attributes [43,44].

Alternatives Criteria

Rock substance strength Fracture spacing Fracture strength

Wa Ma Sa VCa Ca Wa VWa Wa Ma Sa

Open pit mining 3 4 4 2 3 4 4 2 3 4


Block caving 4 1 1 4 4 3 0 4 3 0
Sublevel stoping 49 3 4 0 0 1 4 0 2 4
Sublevel caving 0 3 3 0 2 4 4 0 2 2
Long wall mining 4 1 0 4 4 0 0 4 3 0
Room and pillar 0 3 4 0 1 2 4 0 2 4
Shrinkage stoping 1 3 4 0 1 3 4 0 2 4
Cut and ll 3 2 2 3 3 2 2 3 3 2
Top slicing 2 3 3 1 1 2 4 1 2 4
Stull stoping 4 1 1 4 4 2 1 4 3 2

Rock substance strengthfracture strength: W: weak (<8); M: moderate (815); S: strong (>15); fracture spacing: VC: very close (020); C: close (2140); W: wide (4170);
VW: very wide (71100).
a
Mining methods.

Table 5
Weighting procedure of the Nicholas MMS technique: Hanging wall attributes [43,44].

Alternatives Criteria

Rock substance strength Fracture spacing Fracture strength


a a a a a a a
W M S VC C W VW Wa Ma Sa

Open pit mining 3 4 4 2 3 4 4 2 3 4


Block caving 4 2 1 3 4 3 0 4 2 0
Sublevel stoping 49 3 4 49 0 1 4 0 2 4
Sublevel caving 3 2 1 3 4 3 1 4 2 0
Long wall mining 4 2 0 4 4 3 0 4 2 0
Room and pillar 0 3 4 0 1 2 4 0 2 4
Shrinkage stoping 4 2 1 4 4 3 0 4 2 0
Cut and ll 3 2 2 3 3 2 2 4 3 2
Top slicing 4 2 1 3 3 3 0 4 2 0
Stull stoping 3 2 2 3 3 2 2 4 3 2
a
Mining methods.

Table 6
Weighting procedure of the Nicholas MMS technique: Footwall attributes[43,44].

Alternatives Criteria

Rock substance strength Fracture spacing Fracture strength


a a a a a a a
W M S VC C W VW Wa Ma Sa

Open pit mining 3 4 4 2 3 4 4 2 3 4


Block caving 2 3 3 1 3 3 3 1 3 0
Sublevel stoping 0 2 4 0 0 2 4 0 1 4
Sublevel caving 0 2 4 0 1 3 4 0 2 4
Long wall mining 2 3 3 1 2 4 3 1 3 3
Room and pillar 0 2 4 0 1 3 3 0 3 3
Shrinkage stoping 2 3 3 2 3 3 2 2 2 3
Cut and ll 4 2 2 4 4 2 2 4 4 2
Top slicing 2 3 3 1 3 3 3 1 2 3
Stull stoping 4 2 2 4 4 2 2 4 4 2
a
Mining methods.

Table 7 also using a simple weighting process, he gave the possibility of


Weighting procedure of the Nicholas MMS technique: Weighting Factors [44]. prioritizing the different mining methods based on those criteria to
Criteria Weighting factors DM. In spite of having this advantage, in his technique all of the
criteria were of equal importance [43]. He therefore modied his
Group 1 Group 2 Group 3
technique by introducing weighting factor for different criteria in
Ore geometry 1 1 1 1992 [44]. There were yet some defects including wrong
Ore zone ground 1.33 0.75 1
conditions
denition of weights and small scoring domain between the
Hanging wall 1.33 0.6 0.8 maximum and minimum scores [7]. So, Miller et al. decided to
ground conditions modify the Nicholas technique. In 1995 they represented the
Foot wall ground 1.33 0.38 0.5 University of British Columbia (UBC) method for MMS. In UBC
conditions
method, the scoring domain between the maximum and minimum
A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061 1043

Table 8
HTOM: rst proposed hierarchy model to MMS considering to technical and operational characteristics.

Goal Criteria S-criteria 1 S-criteria 2 Descriptions

Technically and (B1) Technical (C11) Geometry conditions D1c11 Ore body thickness Intervals of upper and lower face of ore body including layer
operationally most parameters or massive Shape
suitable mining D2c11 Ore body shape General solid shape of ore body
method (A) D3c11 Ore body depth An interval of ground level and upper face of ore body
D4c11 Ore body dip Dip of general shape of ore body
D5c11 Grade distributions Type of minerals distribution in the area. It can be existed in
form of erratic, gradational or uniform.
D6c11 Ore body volume Volume of the overall ore reserve
(C12) Geomechanical D1c12 Ore body RMR Structural and mechanical quality of ore body rock mass
conditions D2c12 Hanging wall RMR Structural and mechanical quality of upper face of ore
bodys rock mass which is surrounded by none
mineralization zone
D3c12 Footwall RMR Structural and mechanical quality of lower face of ore body
rock mass which is surrounded by none mineralization
zone
(C13) Geographical D1c13 Hydrogeology conditions Underground hydrogeology level
conditions D2c13 Climate of area Weather conditions of mine
(B2) Operational (C21) production D1c21 Production rate Rate of ore production per mining sequence
parameters D2c21 Recovery Ability of ore extraction which is remained during mining
operations
D3c21 Development production Production of ore during development works before start of
the project
D4c21 Production per man shift Obtained ore in terms of miner in each work sequence
(C22) Mining operations D1c22 Selective mining Ability of planning several stope in a mine
D2c22 Flexibility Ability of changing a mining method to another similar
methods
D3c22 Dilution Mixture of ore and waste during mining operations
D4c22 Development rate Rate of achieving to ore body since start of the project
D5c22 Mineable reserve Part of overall ore body which can be extracted by method
completely
D6c22 Technology availability Ability of automation and mechanization of method
D7c22 Labor availability Skilled miners and availability of them
(C23) Mining aspects D1c23 Environmental aspects Environmental effects of mining in the area
D2c23 Safety Overall safety aspects of method on miners and spaces
D3c23 Occupational interests Ability of more people in area to be occupied by method

scores was extended [40]. But it did not seem a comprehensive little by little getting near to their purpose which was selecting the
method for all of the mines because it paid an extraordinary optimum mining method, none of them was a suitable and perfect
attention to the stoping methods used in the mines of Canada [7]. solution for selection and representation the optimum mining
In addition, the importances of the criteria were not taken into method. In this way a number of studies were also carried out by
consideration [40]. Although the proposed MMS methods were other researchers emphasizing the underground mining methods

Table 9
HEM: Second proposed hierarchy model with economical criteria to select optimum mining method.

Goal Criteria S-criteria Descriptions

Optimum method (B1) costs (C11) Capital costs Include costs of planned mine development,
selection (A) manufacturers quotations of equipments, quantities and
contractors unit costs for constructions, installations and
other services [45]. Such as land ownership, equipment
purchasing, building communication etc.
(C12) Operating costs Include quantities and unit costs of all components of
supplies and labor [45], such as consumption fuel and
energy, drilling rods, explosive agents, salaries, equipment
requirements and repairs, etc.
(C13) Reclamation/ Include costs of reclamation or rehabilitation of mining area
rehabilitation costs after nishing of the mine life. For example, an open pit
mine cone can be changed into a lake.
(B2) Initial investment It is directly depended on mine development duration.
rate of return Return of initial investment in mining methods which have
shorter development cycle is faster than other methods.
(B3) Income (C14) Income per tone It is different in various mining method due to its variety in
of ore overall mining cost.
(C15) Equipment worth Worth of drilling, loading, transporting and other mining
and its usages equipments and their usability in other mining works.
Usability of underground mining equipments, for example
shearer loaders is in similar condition such as long wall
mining only. Even though an open pit hydraulic excavator
can be employed in other mining or construction works
such as trenching.
1044 A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061

Fig. 3. 3D view of Western ore body of north anomaly of Choghart iron mine using SURPAC software. Three main mineralization zones, ground level, outcrop of middle ore
body and other visual characteristics are observable [30].

Table 10
Characteristics of western ore body of north anomaly [30].

Parameters Value Descriptions

Thickness 40 m Average value of thickness has been considered. It is various and has a minimum and
maximum value of 10110 m respectively
Depth 0500 m Due to outcrops, upper bound of ore body is begin over ground level (+20)l and is
extended to average 500 m (500) from ground level
Dip 408 An average value of dip is mentioned for overall shape. Ore body 1 has massive general
shape and an approximately dip should be dened for it
Dip direction Az > 3403558 North-Northwest-North Azimuth of general shape of ore body: angle of longest dip line in a inclined layer ore
Ore body RMR 3040% Structural and mechanical quality of ore body rock mass
Walls RMR 3040% Structural and mechanical quality of walls rock mass. In hanging wall layers which are
nearer than ground level, it is close to 30
Shape Massive (more or less) A massive shape is appeared in 3D views of ore body models
Ore grade %26.56gradational It has an erratic position of grade distribution and has been classied in poor grade
category
Ore body volume 40,000,000 m3 It includes three main ore body and other mineralization zones

Table 11
Process of selecting the suitable mining method by the Nicholas technique.

Mining method Ore Hanging Wall Foot Wall Total


weights
Criteria

General Ore Ore Grade RSS: FSp: FSt.: RSS: FSp: FSt.: RSS: FSp: FSt.:
shape: thickness: plunge distribution: moderate wide moderate moderate wide moderate moderate wide moderate
massive thick intermediate graduate

Block 4 2 2 2 1 3 3 2 3 2 3 3 3 33
caving (BC)
Sublevel 3 4 1 2 3 4 2 2 3 2 2 3 2 33
caving (SLC)
Cut and 0 0 3 3 2 2 3 2 2 3 2 2 4 28
ll (C&F)
Open pit 3 4 3 3 4 4 3 4 4 3 4 4 3 46
mining (OP)
Sublevel 2 4 1 3 3 1 2 3 1 2 2 2 1 27
stoping (SLS)
RSS: rock substance strength; FSp: fracture spacing; FSt: fracture strength.
A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061 1045

Fig. 4. Visualized model of technicaloperational hierarchy analysis.

Table 13
Table 12 Fuzzy pair-wise comparison matrix of sub-criteria (1) (i.t.o. B1).
Fuzzy pair-wise comparison matrix of criteria (i.t.o. A).
B1 C11 C12 C13
A B1 B2
C11 (1/1, 1/1) (5/6, 7/8) (7/8, 9/9)
B1 (1/1, 1/1) (5/6, 7/8) C12 (5/6, 7/8)1 (1/1, 1/1) (1/1, 2/3)
B2 (5/6, 7/8)1 (1/1, 1/1) C13 (7/8, 9/9)1 (1/1, 2/3)1 (1/1, 1/1)
1046 A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061

Table 14
Fuzzy pair-wise comparison matrix of sub-criteria (1) (i.t.o. B2).

B2 C21 C22 C23

C21 (1/1, 1/1) (2/3, 4/5) (5/6, 7/8)


C22 (2/3, 4/5)1 (1/1, 1/1) (1/2, 3/4)
C23 (5/6, 7/8)1 (1/2, 3/4)1 (1/1, 1/1)

Table 15
Fuzzy pair-wise comparison matrix of sub-criteria (2) (i.t.o. C13).

C12 D1c12 D2c12 D3c12


1
D1c12 (1/1, 1/1) (2/3, 3/4) (2/2, 2/3)
D2c12 (2/3, 3/4) (1/1, 1/1) (6/6, 7/8)
D3c12 (2/2, 2/3)1 (6/6, 7/8)1 (1/1, 1/1) Fig. 5. Trapezoidal fuzzy number of alternatives nal weights based on technical
and operational considerations.

Table 16 Yavuz applied three models of data base management, model


Fuzzy pair-wise comparison matrix of sub-criteria (2) (i.t.o. C13).
management and dialog subsystem. His methods which was called
C13 D1C 13 D2C 13 DDS, was applicable for resolving the multi-criteria decision
D1C 13 (1/1, 1/1) (3/4, 5/6) making problems of mines, using crisp numbers [3]. But this
D2C 13 (3/4, 5/6)1 (1/1, 1/1) system, like some of the preceding proposed systems, was
applicable for a certain category of mining methods. In other
words, a mining engineer was able to use DDS for selecting just a
Table 17 single underground mining method. Almeida et al. carried out a
Fuzzy pair-wise comparison matrix of sub-criteria (2) (i.t.o. C23).
research leading to representation a solution for MMS problem. In
C23 D1c23 D2c23 D3c23 this study PROMETHEE technique was used in order to selecting a
D1c23 (1/1, 1/1) (1/2, 2/3) (5/6,6/7) mining method for mineral resources [2]. This method, as opposed
D2c23 (1/2, 2/3)1 (1/1, 1/1) (2/3,3/4) to the preceding ones, covered all of the surface and underground
D3c23 (5/6, 6/7)1 (2/3, 3/4)1 (1/1,1/1) mining methods. For the user of such a method, allocation of
engineering judgments was only possible by using crisp numbers.
In 2003, Osanloo et al. used conventional AHP in order to nd out a
[6,47,53]. In 2003, Guaray et al. proposed a compound method suitable mining method in anomaly No. 3 of GOL-E-GOHAR mine in
based on expert systems and knowledge base, which was usable Iran. In this study, two hierarchical models were used including
for underground mining methods [24]. In a similar research, in some of the technical an economical criteria, respectively [46]. This
order to representing a solution for MMS problem, Alpay and method used crisp numbers in order to allocating judgments. In

Table 18
Fuzzy pair-wise comparison matrix of sub-criteria (2) (i.t.o. C11).

C11 D1c11 D2c11 D3c11 D4c11 D5c11 D6c11


1
D1c11 (1/1, 1/1) (2/2, 3/4) (2/2, 3/4) (1/1, 1/2) (3/3, 4/5) (3/3,4/5)
D2c11 (2/2, 3/4)1 (1/1, 1/1) (5/6, 7/7)1 (1/2, 3/3)1 (1/2, 2/3) (1/1, 2/3)
D3c11 (2/2, 3/4) (5/6, 7/7) (1/1, 1/1) (1/2, 2/3) (8/9, 9/9) (6/7, 8/8)
D4c11 (1/1, 1/2)1 (1/2, 3/3) (1/2, 2/3)1 (1/1, 1/1) (2/4, 4/4) (2/3, 4/5)
D5c11 (3/3, 4/5)1 (1/2, 2/3)1 (8/9, 9/9)1 (3/4, 4/4)1 (1/1, 1/1) (1/1, 2/2)1
D2c21 (3/3, 4/5)1 (1/1, 2/3)1 (6/7, 8/8)1 (2/3, 4/5)1 (1/1, 2/2) (1/1, 1/1)

Table 19
Fuzzy pair-wise comparison matrix of sub-criteria (2) (i.t.o. C22).

C22 D1c22 D2c22 D3c22 D4c22 D5c22 D6c22 D7c22

D1c22 (1/1, 1/1) (1/2, 2/3) (1/1, 2/2)1 (1/1, 1/2)1 (2/3, 4/5)1 (1/2, 3/4)1 (1/1, 2/3)1
D2c22 (1/2, 2/3)1 (1/1, 1/1) (2/3, 4/4)1 (2/2, 2/3)1 (4/5, 6/7)1 (4/5, 6/7)1 (1/2, 2/3)1
D3c22 (1/1, 2/2) (2/3, 4/4) (1/1, 1/1) (1/1, /2/2) (1/2, 3/4)1 (1/2, 3/3)1 (1/1, 2/3)
D4c22 (1/1, 1/2) (2/2, 2/3) (1/1, /2/2)1 (1/1, 1/1) (3/4, 5/6)1 (2/3, 4/5)1 (1/1, 2/3)1
D5c22 (2/3, 4/5) (4/5, 6/7) (1/2, 3/4) (3/4, 5/6) (1/1, 1/1) (1/1, 2/3) (3/3, 3/4)
D6c22 (1/2, 3/4) (4/5, 6/7) (1/2, 3/3) (2/3, 4/5) (1/1, 2/3)1 (1/1, 1/1) (3/3, 3/4)
D7c22 (1/1, 2/3) (1/2, 2/3) (1/1, 2/3)1 (1/1, 2/3) (3/3, 3/4)1 (3/3, 3/4)1 (1/1, 1/1)

Table 20
Fuzzy pair-wise comparison matrix of sub-criteria (2) (i.t.o. C21).

C21 D3c22 D2c21 D3c21 D4c21

D1c21 (1/1, 1/1) (2/3, 4/5) (3/4, 4/5) (1/1, 1/2)


D5c22 (2/3, 4/5)1 (1/1, 1/1) (1/1, 1/2) (2/3, 4/5)1
D3c21 (3/4, 4/51 (1/1, 1/2)1 (1/1, 1/1) (2/3, 4/5)1
D4c21 (1/1, 1/2)1 (2/3, 4/5) (2/3, 4/5) (1/1, 1/1)
A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061 1047

Fig. 6. Inuence of each sub-criteria of level four on each alternative in HTOM.

Fig. 7. Visualized model of economical hierarchy analysis.

addition, as the method was only limited to the aforementioned coal basin [60]. In 2008, Naghadehi et al., using FAHP, represented a
mine, the selected criteria were insufcient. In 2004, Bitarafan and solution for MMS for a Bauxite mine in Iran. They made use of
Ataei, used fuzzy MCDM in order to resolving MMS problem in the thirteen different criteria [42]. Their study was just carried out for
aforementioned mine. Their method was based on Yager method. underground mining method.
They used fteen criteria which were taken into account, In the above-mentioned literature, shortcomings such as
considering ore geometry, rock quality and its strength [7]. In linguistic and unreliable denitions for criteria, structure of
this study, the ultimate prioritizing of the mining methods was weighting system of some of the methods which are based
only based on technicaloperational considerations, without an on weighting, a lack of fuzzy environment, using insufcient
economical viewpoint. In other words the ultimately selected the technicaloperational criteria, and a subjective view toward some
mining method based on their study might be uneconomical. of the represented solutions led to problems such as inability for
Also in 2004, Xinchun et al. resolved MMS problem by using
FAHP and criteria which were taken into account for a certain

Fig. 8. Trapezoidal fuzzy number of three choice of mining in HEM. Fig. 9. Inuence of sub-criteria on each alternative of HEM.
1048 A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061

selecting a suitable mining method or selecting a mining number and other types of it. A general type trapezoidal fuzzy
method which did not take economical considerations into number is illustrated in Fig. 1.
account. In addition some of these solutions were designed just The function which represents a grade of membership
for selecting an underground mining method. It is worth for numbers is called membership function. A fuzzy number is
mentioning that in the previous studies, there was not the a special kind of fuzzy set where F fx; mx; x 2 Rg; R : 1
possibility of mathematical modeling of the complicated and < x < 1; mx : R0; 1. So, a trapezoidal fuzzy number is denoted
uncertain concept of MMS. It is therefore obvious that MMS as A a=b; g =d which has a value of membership 0 for x  a and
concept which includes a purpose, some criteria, sub-criteria x  d, (x  a)/(b  a) for a  x  b, 1 for b  x  g and (d  x)/
and alternatives, is a decision making problem, one of the best (d  g) for g < x  d.
possible solutions of which is using decision making tools. In fact Thus considering Fig. 1, membership value of b and c is 1. If
Nicholas resolved the decision making MMS problem by a 2 (0,1], then alpha cut of A, a fuzzy subset of universal set X,
representing a numerical system which was a simple decision written as A[a], is dened to be the crisp set {xjA(x)  a}. This set is
making model just for mining. In his model, the surface and the collection of all x in X whose membership value is at least alpha.
underground mining methods were used as alternatives. It is also In this paper, we named a x value of alpha level cut 0.5 by j. So, in a
worth noting that only in his technique the most protable mining trapezoidal fuzzy number, jL is a crisp value of A(x) which is
method is emphasized. So, in this paper by using the idea dened between value of a, b. A value of jR also represents crisp
presented in Nicholas technique, it has been tried to modify it and value of alpha level cut 0.5 between g, d. Now the fuzzy arithmetic
also to represent a new approach to MMS problem in mines so that can be described as follows. If A a1 ; b1 ; g 1 ; d1 > 0 and zero does
its deciencies such as weighting problems would be eliminated. not belong to the support of B a2 ; b2 ; g 2 ; d2 , the main
As a result, considering the structural similarity between Nicholas arithmetic of two fuzzy numbers denes as below [52]:
technique and hierarchical analysis process, AHP has been used as a a
the framework of proposed method for weighting. Considering the A  B aa1  b1 ; aa2  b2  (1)
inability of crisp numbers to assign the viewpoint of DM to
a a
uncertainties existing in mines resources, fuzzy sets theory was A  B aa1  b1 ; aa2  b2  (2)
used to compensate for this inability. Two hierarchical models of
" #
HTOM and HEM are therefore developed in order for hierarchical 1 1
analysis. These models are based on technicaloperational and A  B aa1 ; aa2   a; a (3)
b1 b2
economical criteria respectively. In HTOM, mining alternatives
which are the same as mining methods are prioritized. The most Considering Eqs. (1)(3) fuzzy addition, subtraction, multi-
protable of them is then selected by HEM. plication and division of two above-mentioned trapezoidal fuzzy
In the following, the fuzzy sets theory and its applications in numbers are dened as follows, respectively [52]:
mining sciences are reviewed. Section 3 describes AHP and its
extensions to fuzzy environment. Section 4 reviews the Nicholas A B a1 a2 =b1 b2 ; c1 c2 =d d2  (4)
methodology in order to know the basic framework of proposed
approach. Section 4.2 explains the proposed approach and its AQB a1  a2 =b1  b2 ; c1  c2 =d  d2  (5)
elements. As a case study, applying the Nicholas technique and the
proposed approach to a mineral resource is illustrated in Section 5. A
B a1  a2 =b1  b2 ; c1  c2 =d1  d2  Cm (6)
A comparison study between those two MMS methodology is also
represented in Section 5.3. In Section 6, we discuss about results of  
a1 b 1 c 1 d1
case study. Finally, section 7 is dedicated to describe the results of AfB = ; = Cd (7)
d2 c2 b2 a2
this study 
1
Reciprocal concept of A also is dened as A 1d = g1 ; b1 = a1
2. Fuzzy sets and fuzzy numbers Cr . We presented results of above-mentioned Eqs. in form which

Fuzzy sets theory was introduced by Zadeh to deal with the


conception of the uncertainty due to imprecision and vagueness
[62]. It provides a strict mathematical framework in which vague
conceptual phenomena can be precisely studied. It can be con-
sidered as a suitable modeling language for vague and imprecision
conceptual relations, criteria and phenomena. Imprecision here is
meant in the sense of vagueness than the lack of knowledge about
the value of a parameter [64]. In fact, it is a natural way to deal with
imprecision problems by denition of class which is representing
continuum grades of membership. Therefore a much wider scope of
applicability in the eld of pattern classication and information
process is accessible by fuzzy sets [62].
A fuzzy set A can be dened as follows: if X is a collection of
objects denoted generically by x, then A is a set of ordered pairs:
A fx; mA xjx 2 Xg. mA x is called the membership function
which is address a degree of compatibility of x in A that maps X to
the membership space M with a range of nite and nonnegative
numbers [62,64]. Fuzzy numbers represent a number of whose
value we are somewhat uncertain. Though they are used for nite
area conventionally, they have special shapes such as triangular
(piecewise linear), s-shape (piecewise quadratic), bell-shaped and Fig. 10. Schematic model of overall progress for selecting a practical and protable
trapezoidal [52]. In this paper, we employed trapezoidal fuzzy mining method in north anomaly of Choghart iron mine.
A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061 1049

is used by Buckley. We also use symbol to describe conception model is weighted using the comparison matrices. Then, the
of approximately equal to. So, symbol of / is not division weight of each subsection of alternatives level must be multiplied
sign. It is applied to separate value of a, b and g, d in a trapezoidal by the weight of all subsections of the upper level. Multiplication of
fuzzy number [11,12]. It should be noted that Cm , Cd , Cr are not a the weight of the subsections of each level must be carried out
trapezoidal fuzzy number. Results of them are approximately considering the path depicted from each subsection to the other.
equal to their own real values and graph. Using a a 2 [0, 1] level The nal weight of each alternative is obtained from the addition of
cut for example a = 0.5, lead to the graph of Cm , Cd , Cr be closed to the weights of the subsections of each path.
their own real state and difference between these two state is been So, consistency of matrices should be also checked by DM.
often negligible. Consistency index of hierarchy (CI) is obtained by CI = (lmax  n)/
n  1, where n is the matrix size, lmax is a principal eigenvalue of
2.1. Fuzzy sets theory in mining sciences: reasons and applications matrix and always lmax  n. Furthermore, a ratio of CI has been
dened to check the judgment consistency as the consistency ratio
Most stages of mining operations deal with geo materials (CR) which is shown in Table 2. The CR is acceptable, if it does less
(rock and soil). These materials usually are anisotropic and than 10%. Otherwise, the comparison matrix is inconsistent and
inhomogeneous. It means that geo materials do not have the same DMs have to redo it to improve. So, judgments should be reviewed
properties in different directions and locations. So, their by DM. Based on consistency theorem of AHP, a pair-wise matrix is
behaviors such as deformability modulus, strength, brittleness, consistent if and only if lmax = n. In term of reciprocal values it is
permeability and discontinuity frequency cannot be realized important to note that if priority of i respect to j be a 2 [1,9], then
precisely [27]. In mine resources, these geo materials (including priority of j respect to i is shown as a1 = 1/a [4,49,50].
ore and waste) also do not have a certain state in their geometric,
geological and mechanical characteristics. Assigning the crisp 3.2. Using AHP in fuzzy environment
numbers to these characteristics will therefore be complicated
and imprecise. In order to model the uncertainty and imprecision As discussed in the previous section, in conventional AHP, DM
in geo materials, it is therefore appropriate to use the fuzzy judgments are carried out using just crisp numbers. Even though
numbers. Although the computing process will be rather huge, the purpose of AHP is to capture the experts knowledge, the
using fuzzy numbers instead of crisp numbers, the uncertain state conventional AHP cannot reect the human thinking style [54]. In
of geo materials will be modeled better. Fuzzy sets theory has a practice, DM encounters much ambiguities and uncertainties in
wide range of application in many parts of mining sciences (see decision making problems. Due to inability of crisp numbers to
[36,22]), such as decision making in mining projects (except in the assign the thinking style and judgment of human being to these
case of applications of fuzzy MCDM tools in MMS problems ambiguities and uncertainties, DM, while using conventional AHP
mentioned in previous section, see also [31]), mineral processing, for assigning their judgments, encounter difculty. In other words,
blasting, rock mechanics, underground mining, tunneling and it is usually the case that DM, by using crisp numbers, cannot
underground spaces studies, mineralogy, reserve evaluating, assign his judgments as accurately as using fuzzy numbers. Using
rock slope stability, acid mine drainage, engineering geology, AHP in fuzzy environment is therefore a suitable solution for the
subsidence in underground mines, etc. [36]. complicated problems of decision making which include ambi-
guity and uncertainty [21,38].
3. Multi-criteria decision making (MCDM) During the rst years of 1980s, the rst methods of FAHP were
represented [56], while conventional AHP and the rst numerical
Although MCDM tools are applied for resolving a wide range of technique for MMS had been introduced simultaneously [43,50].
engineering problems such as planning, evolution, forecasting, etc., Representation of the different methods of FAHP has been
it is commonly used as a solution for decision making problems continued up to now. The characteristics of these methods are
[23,48]. In other words, in order to nd out the most suitable briey described in Table 58 in Appendix A. Some of the recent
alternative from among a number of possible ones, MCDM studies have revealed the applications of FAHP [18,29,32,34,51,55].
is applied based on a number of criteria. This selecting process Moreover, several researchers have also been discussed about the
is carried out in different forms such as outranking, prioritizing fundamental concepts of FAHP [20,35,57,58,63].
and distance from ideal solution [23].
3.3. Buckleys FAHP method
3.1. Decision making using AHP
In case of weighting process, we describe the geometric mean
In 1998, Saaty introduced AHP as one of the MCDM tools. Using method of conventional AHP to develop it into fuzzy environment.
AHP, DM is able to prioritize and select an alternative among a Suppose a comparison matrix with crisp elements [A] as below
number of possible ones based on his experience and knowledge [50]:
[50]. For this purpose, in order to construct a hierarchical model 2 3
which includes three main levels of target, criteria and alternatives, a1;1 a1;2 a1;m
6 a2;1 a2;2 a2;m 7
DM can compare the subsections of each level to those of the upper 6 7
A6 . .. .. .. 7 (8)
level. In case of existing a sub-criteria level in the hierarchical 4 .. . . . 5
model, each one of those sub-criteria will be compared to each am;1 am;2 am;m
other, based on that criterion from which the sub-criteria are
Geometric mean of each row calculates by Zi where:
derived. In such a case, the alternatives are compared to each other
based on each subsection of the sub-criteria level located on the 0 11=m
upper level. The process of comparison is based on construction a Y
m
ri @ ai j A (9)
comparison pair-wise matrix (n  n) for each level, n being the j1
number of subsections. Crisp numbers represented in Table 1 must
be used to allocate the amount of importance of each subsection of Now crisp weight of element i is obtained by (10):
a hierarchical level to the other subsections of the same level. After
weighting process, the subsection of each level of the hierarchical wi r i  r 1 r 2 r m 1 (10)
1050 A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061

Table 21
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D1c11 ).

D1c11 BC SLC C&F OP SLS


1
BC (1/1, 1/1) (3/3, 4/6) (5/6, 7/8) (1/1, 2/3) (5/6, 6/7)
SLC (3/3, 4/6)1 (1/1, 1/1) (1/1, 2/3) (4/5, 6/6)1 (1/1, 2/3)
C&F (5/6, 7/8)1 (1/2, 2/3)1 (1/1, 1/1) (6/6, 7/8)1 (1/1, 2/3)1
OP (1/1, 2/3) (4/5, 5/6) (7/8, 8/9) (1/1, 1/1) (7/8, 8/9)
SLS (3/4, 5/6)1 (1/2, 2/3)1 (2/3, 3/4) (4/5, 5/6)1 (1/1, 1/1)

Table 22
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D2c11 ).

D2c11 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/2, 3/3)1 (2/3, 3/4)1 (3/3, 4/5)1 (1/1, 2/3)1
SLC (1/2, 3/3) (1/1, 1/1) (1/1, 2/3)1 (2/2, 2/3)1 (1/1, 2/3)
C&F (2/3, 3/4) (1/1, 2/3) (1/1, 1/1) (1/1, 2/3)1 (1/1, 2/3)
OP (3/3, 4/5) (2/2, 2/3) (1/1, 2/3) (1/1, 1/1) (3/3, 4/5)
SLS (1/1, 2/3) (1/1, 2/3)1 (1/1, 2/3)1 (3/3, 4/5)1 (1/1, 1/1)

Table 23
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D4c11 ).

D4c11 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/1, 1/2)1 (1/2, 2/3)1 (3/3, 4/5)1 (2/2, 3/4)
SLC (1/1, 1/2) (1/1, 1/1) (1/2, 2/3)1 (3/4, 5/6)1 (1/1, 2/3)
C&F (1/2, 2/3) (1/2, 2/3) (1/1, 1/1) (1/1, 2/3)1 (4/4, 5/6)
OP (3/3, 4/5) (3/4, 5/6) (1/1, 2/3) (1/1, 1/1) (5/6, 7/7)
SLS (2/2, 3/4)1 (1/1, 2/3)1 (4/4, 5/6)1 (5/6, 7/7)1 (1/1, 1/1)

Table 24
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D3c11 ).

D3c11 BC SLC C&F OP SLS

BC (1/1, 1/1) (2/3, 3/4) (3/3, 4/5) (1/2, 3/4)1 (3/4, 4/5)
SLC (2/3, 3/4)1 (1/1, 1/1) (1/1, 2/3) (5/6, 7/7)1 (1/1, 2/3)
C&F (3/3, 4/5)1 (1/1, 2/3)1 (1/1, 1/1) (7/8, 8/9)1 (1/1, 2/2)
OP (1/2, 3/4) (5/6, 7/7) (7/8, 8/9) (1/1, 1/1) (7/8, 9/9)
SLS (3/4, 4/5)1 (1/1, 2/3)1 (1/1, 2/2)1 (7/8, 9/9)1 (1/1, 1/1)

Table 25
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D5c11 ).

D5c11 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/3, 3/3) (3/4, 4/5) (1/1, 2/3)1 (3/4, 4/4)
SLC (1/3, 3/3)1 (1/1, 1/1) (1/1, 2/2) (5/6, 6/7)1 (1/1, 1/2)
C&F (3/4, 4/5)1 (1/1, 2/2)1 (1/1, 1/1) (5/6, 7/8)1 (1/1, 2/3)
OP (1/1, 2/3) (5/6, 6/7) (5/6, 7/8) (1/1, 1/1) (5/6, 7/7)
SLS (3/4, 4/4)1 (1/1, 1/2)1 (1/1, 2/3)1 (5/6, 7/7)1 (1/1, 1/1)

Table 26
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D6c11 ).

D6c11 BC SLC C&F OP SLS

BC (1/1, 1/1) (3/4, 5/6) (5/6, 6/6) (1/2, 2/3)1 (3/4, 5/6)
SLC (3/4, 5/6)1 (1/1, 1/1) (1/1, 2/3) (4/5, 6/6)1 (1/1, 1/2)1
C&F (5/6, 6/6)1 (1/1, 2/3)1 (1/1, 1/1) (6/7, 7/8)1 (1/1, 2/3)1
OP (1/2, 2/3) (5/6, 6/6) (6/7, 7/8) (1/1, 1/1) (5/6, 6/6)
SLS (3/4, 5/6)1 (1/1, 1/2) (1/1, 2/3) (7/7, 8/8)1 (1/1, 1/1)

If A is consistent then geometric mean method always produces fuzzy trapezoidal number in the gure of (a/b, g/d) where
equal weights as Saatys lmax method and so, if m = 3 then both 0  a  b  g  d 2 S. Because of their easiness intuition and
methods compute equal weights. For m > 3 results of these two various interpretations in practice, it is easy to construct a fuzzy
methods are close to each other [12]. comparison matrix by fuzzy numbers. But priorities of DM should
Now, we describe how to develop aforementioned method into be represented as a positive reciprocal matrix. A matrix (m  m)
fuzzy environment by Buckleys method. Buckley represented a whose entries are ratios a ji a1
i j is called positive reciprocal
A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061 1051

Table 27
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D1c12 ).

D1c12 BC SLC C&F OP SLS


1
BC (1/1, 1/1) (2/2, 3/4) (3/4, 4/5) (1/2, 2/3) (2/3, 4/4)
SLC (2/2, 3/4)1 (1/1, 1/1) (1/2, 2/3) (4/5, 6/7)1 (1/1, 1/2)
C&F (3/4, 4/5)1 (1/2, 2/3)1 (1/1, 1/1) (5/6, 6/7)1 (1/1, 2/2)1
OP (1/2, 2/3) (4/5, 6/7) (5/6, 6/7) (1/1, 1/1) (4/5, 6/7)
SLS (2/3, 4/4)1 (1/1, 1/2)1 (1/1, 2/2) (4/5, 6/7)1 (1/1, 1/1)

Table 28
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D2c12 ).

D2c12 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/1, 1/2)1 (2/2, 2/3)1 (3/4, 4/5)1 (2/3, 3/4)1
SLC (1/1, 1/2) (1/1, 1/1) (2/2, 2/3)1 (3/4, 4/5)1 (1/2, 3/4)1
C&F (2/2, 2/3) (2/2, 2/3) (1/1, 1/1) (2/2, 3/4)1 (1/1, 1/2)
OP (3/4, 4/5) (3/4, 4/5) (2/2, 3/4) (1/1, 1/1) (2/2, 3/3)
SLS (2/3, 3/4) (1/2, 3/4) (1/1, 1/2)1 (2/2, 3/3)1 (1/1, 1/1)

Table 29
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D3c12 ).

D3c12 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/1, 1/2)1 (1/1, 2/3) (6/7, 8/9)1 (2/2, 2/3)1
SLC (1/1, 1/2) (1/1, 1/1) (2/2, 2/3) (6/7, 7/8)1 (1/2, 3/4)1
C&F (1/1, 2/3)1 (2/2, 2/3)1 (1/1, 1/1) (7/7, 8/9)1 (1/2, 2/3)1
OP (6/7, 8/9) (6/7, 8/9) (7/7, 8/9) (1/1, 1/1) (2/3, 4/5)
SLS (2/2, 2/3) (1/2, 3/4) (1/2, 2/3) (2/3, 4/5)1 (1/1, 1/1)

Table 30
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D2c13 ).

D2c13 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/1, 2/2) (1/2, 2/3) (1/1, 1/2) (1/1, 1/2)
SLC (1/1, 2/2)1 (1/1, 1/1) (1/1, 2/2) (1/1, 2/2) (1/1, 2/2)
C&F (1/2, 2/3)1 (1/1, 2/2)1 (1/1, 1/1) (1/2, 2/3) (1/1, 1/2)1
OP (1/1, 1/2)1 (1/1, 2/2)1 (1/2, 2/3)1 (1/1, 1/1) (1/1, 2/3)1
SLS (1/1, 1/2)1 (1/1, 2/2)1 (1/1, 1/2) (1/1, 2/3) (1/1, 1/1)

Table 31
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D1c13 ).

D1c13 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/2, 2/3) (3/4, 4/5) (1/2, 3/4)1 (2/3, 4/5)
SLC (1/2, 2/3)1 (1/1, 1/1) (1/2, 2/3) (4/4, 5/6)1 (1/2, 3/4)
C&F (3/4, 4/5)1 (1/2, 2/3)1 (1/1, 1/1) (7/8, 8/8)1 (1/1, 1/2)1
OP (1/2, 3/4) (4/4, 5/6) (7/8, 8/8) (1/1, 1/1) (7/8, 8/9)
SLS (2/3, 4/5)1 (1/2, 3/4)1 (1/1, 1/2) (7/8, 8/9)1 (1/1, 1/1)

Table 32
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D1c21 ).

D1c21 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/2, 2/3) (4/5, 5/5) (1/1, 2/3)1 (1/2, 2/3)
SLC (1/2, 2/3)1 (1/1, 1/1) (2/2, 3/4) (2/3, 4/5)1 (1/1, 1/2)
C&F (4/5, 5/5)1 (2/2, 3/4)1 (1/1, 1/1) (4/5, 6/7)1 (2/2, 3/4)1
OP (1/1, 2/3) (2/3, 4/5) (4/5, 6/7) (1/1, 1/1) (2/3, 4/5)
SLS (1/2, 2/3)1 (1/1, 1/2)1 (2/2, 3/4) (2/3, 4/5)1 (1/1, 1/1)

matrix [11]. Therefore DMs can employ all eight types of them priority of element ith to jth is approximately for b = g equal to 1.
to express their judgments easily. Except for real number So, a = b and g = d explain importance between a = b to 1 and g = d
a = b = g = d, and also for a normal shape of trapezoidal number to 1. Importance least a to 1 is explaining a priority number
which has been shown in Fig. 1, other six types of fuzzy hierarchy which is a = b = g. Otherwise a normal form of trapezoidal fuzzy
numbers is shown in Fig. 2. In fuzzy hierarchy analysis, when b = g, number is present as (a/b, g/d) which is explaining importance of b
a triangular shape of fuzzy numbers is formed. It means that to 1 and g to 1.
1052 A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061

Denition. Let A ai j  is a fuzzy positive reciprocal matrix. Now A each row of A ai j , then it is dened as below:
is consistent if and only if aik
ak j ai j [12]. ri ai;1
ai;2

ai;m 1=m (12)
It means that product of multiplication of aik and ak j is
approximately equal to ai j . If A is not consistent, then fuzzy Also, considering (10), fuzzy weight wi is:
judgments should be reviewed and improved by DM. It should be wi ri
r1 r2 rm 1 (13)
noted that improving the consistency means that ratio estimates in
matrix are closer to being logically related than to being randomly For 0  y  1, fi(y) and gi(y) are dened as left and right sides of
chosen. ai j to determine the membership function of wi where:
2 31/m
Theorem 1. Let A ai j  where ai j ai j =bi j ; g i j =di j and let bi j  Y
m

ai j  g i j for all i,j. If A = [aij] is consistent, then A is consistent [12]. f i y 4 bi j  ai j y ai j 5 (14)


j1
2 31/m
In the other words, we can say if bi j ; g i j   bik ; g ik  bk j ; g k j  for Ym
all i,j and k, then A ai j  is perfectly consistent. Moreover, g i y 4 g i j  di j y di j 5 (15)
if consistency of A be reasonable, its consistency is also acceptable. j1
2 31/m
It means that if product of bik ; g ik :bk j ; g k j  also be reasonably Y m

closed to bi j ; g i j , then it is consistent. For example 1=2 2 = 3=5; 1, ai is dened as ai 4 ai j 5 so, a will be obtained by combination
of all ai as below: j1
but since 1/2 is reasonably closed to 3/5, we can conclude that
aik ak j is reasonably close to ai j . Thus A is reasonably consistent X
m
[11]. In the following, Buckleys method has been described as a a ai (16)
i1
FAHP prescription in three steps.
In similar way, bi and b, gi and g, di and d will be dened. Finally,
Step 1: After constructing a hierarchy model of problem, in this X
m X
m
situation, fuzzy comparison matrix should be represented by let f y f i y, gy g i y. So, wi will be obtained as
DM via trapezoidal fuzzy numbers as elements of fuzzy positive i1 i1

reciprocal (m  m) matrix A. In case of reciprocal matrices, if follows:


 
priority of element ith to jth, is ai j ai j =bi j ; g i j =di j , then 1 1
wi ai d =bi g 1 ; g i b =di a1 (17)
priority of jth to ith be dened as (11):
1 1
a ji di j =g 1 1
i j ; bi j =ai j (11) And its membership function mwi is:
8 1
>
> 0 if x  ai d or x  di a1
Whenever elements i and j have equally importance, then >
< 1 1
1 if bi g  x  g i b
ai j a ji 1; 1; 1; 1. mwi 1 (18)
> 1
ai d  x  bi g
Step 2: Now, fuzzy weights of comparison matrices should be > y 2 0; 1 if
>
: 1
computed by geometric mean method. If ri is geometric mean of y 2 0; 1 if g i b  x  di a1

Table 33
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D2c21 ).

D2c21 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/1, 1/2) (1/2, 2/3)1 (1/2, 2/3)1 (1/2, 2/3)
SLC (1/1, 1/2)1 (1/1, 1/1) (1/2, 3/4)1 (2/2, 3/4)1 (1/2, 2/3)
C&F (1/2, 2/3) (1/2, 3/4) (1/1, 1/1) (1/1, 2/3)1 (2/3, 4/4)
OP (1/2, 2/3) (2/2, 3/4) (1/1, 2/3) (1/1, 1/1) (3/4, 4/4)
SLS (1/2, 2/3)1 (1/2, 2/3)1 (2/3, 4/4)1 (3/4, 4/4)1 (1/1, 1/1)

Table 34
Fuzzy pair-wise comparison matrix of alternatives(i.t.o. D3c21 ).

D3c21 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/2, 2/3)1 (1/2, 2/3) (4/4, 5/6)1 (3/4, 4/4)1
SLC (1/2, 2/3) (1/1, 1/1) (3/3, 4/4) (2/2, 3/4)1 (1/2, 2/3)1
C&F (1/2, 2/3)1 (3/4, 4/4)1 (1/1, 1/1) (6/7, 8/9)1 (6/6, 7/7)1
OP (4/4, 5/6) (2/2, 3/4) (6/7, 8/9) (1/1, 1/1) (1/1, 2/3)
SLS (3/4, 4/4) (1/2, 2/3) (6/6, 7/7) (1/1, 2/3)1 (1/1, 1/1)

Table 35
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D4c21 ).

D4c21 BC SLC C&F OP SLS


1
BC (1/1, 1/1) (1/2, 3/3) (3/3, 4/5) (2/3, 3/3) (1/2, 3/4)
SLC (1/2, 2/3)1 (1/1, 1/1) (1/1, 2/3) (5/6, 7/7)1 (1/1, 2/3)
C&F (3/4, 4/5)1 (1/1, 2/3)1 (1/1, 1/1) (6/7, 8/8)1 (1/1, 2/2)1
OP (2/2, 3/3) (5/6, 7/7) (6/7, 8/8) (1/1, 1/1) (5/6, 6/7)
SLS (1/2, 3/4)1 (1/1, 2/3)1 (1/1, 2/2) (5/6, 6/7)1 (1/1, 1/1)
A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061 1053

Considering (18), graph mwi can be drawn from zero to the left 4.2. Proposed approach
1 f i y 1 1
of aid ,x g i y on the interval ai d ; bi g  with a horizontal line
The shortcomings of Nicholas technique are due mainly to the
1 1 g i y 1
from bi g ; 1 to g i b ; 1, x f i y
on the interval g i b ; di a1 , lack of a perfect methodology for weighting. In other words his
and zero to the right of di a1 [12]. We strongly remind that / technique is not able to obtain the weight of each mining method
symbol is not division sign. based on inconsistency ratio of assigned scores. So in proposed
Step 3: As before, using combination of obtained weights approach, by replacing Buckleys FAHP with the simple weighting
corresponding to their own path in hierarchy model, a fuzzy system of Nicholas, the mentioned shortcoming has been
nal weight Wi should be calculated considering to (19): eliminated.
Another defect of his technique is a lack of inuential and
X
m
Wi w j ri j (19) important criteria in MMS process. It has been tried to modify and
j1 complete the criteria used in Nicholas technique so that the mining
alternatives be prioritized based on important technicalopera-
w j , ri j is weight of each alternatives and attributes respectively. tional criteria. For this aim the hierarchical model of HTOM has
So, centroid defuzzication method is preferred to represent the been devised. This model is represented in Table 8. HTOM is a
obtained nal fuzzy weights in form of the crisp numbers by completed version of Tables 36 related to Nicholas technique.
Eqs. (20) [52]. HTOM has two main technical and operational criteria. Each one of
R
xmAi xdx these criteria includes three sub-criteria, each of which is

X R (20) subdivided to smaller elements. By using HTOM, DM is able to
mAi xdx
consider the mining methods as hierarchical analysis alternatives
and to compare them with each other in order to prioritize and
4. MMS procedure select the top three alternatives.
After prioritizing the mentioned alternatives, three mining
4.1. A review of MMS technique of Nicholas alternatives which have the most technicaloperational weight
must be considered as the alternatives of the second hierarchical
Nicholas technique is a simple decision making tool which is model of HEM which is constructed based on economical criteria.
only used in order to select the mining method. This technique In HEM, these three mining alternatives are compared to each
enjoys a simple weighting system, so the way of its application is other again and the most protable of them is nally selected as
not complicated. In addition, in order to assign the scores to mining the suitable method from both technicaloperational and eco-
alternatives, it uses a small range of crisp numbers. In this nomical points of view. HEM is divided into three criteria: costs,
technique thirteen mining parameters are taken into account as initial investment rate of return, and incomes. These criteria,
criteria. Each mining method is weighted based on these criteria except for the second one, include sub-criteria. HEM is shown in
each of which having some sub-criteria, so that based on each sub Table 9. Criteria and sub-criteria related to HTOM and HEM are
criterion, scoring is assigned for each mining method. The ultimate considered based on engineering experiences and feasibility
weight of each mining method is obtained by adding those scores. studies carried out for different mining methods in different
In this way, all of the mining methods are prioritized .in this mines. It must be noted that in compensatory MCDM methods
technique the numbers of 0, 1, 2, 3, 4, and 49 are used in order to such as AHP [37] poor score of a number of criteria can be
weight. Thus the importance of a mining method for a certain compensated by high score in the other criteria. It may be
criterion is indicated by numbers 3 and/or 4. The numbers of 1 and mitigated the effect of poor criteria in total weight of an alternative
2 indicate that the related mining method is probably suitable [28]. So, two separate hierarchical models of HTOM and HEM have
based on related criteria. In the case of number 0, a mining method been constructed. In other words, the two models are considered
is not rejected but it means that the related method is not separately in order to prevent the poor weights of some of the
applicable for that certain criterion. The most weight which may be alternatives in HEM from being compensated by strong weights of
obtained by a mining method is +48. The score of 49 is therefore the same alternatives in HTOM. In this way, after being veried
used in order to reject a mining method which cannot be used for a both technicaloperationally and economically, the optimum
specic condition of resource so that, the method would not have alternative can be ultimately selected.
any chance for being selected [43]. It should be noted that by combining the two models, a trade off
The equality of importance for criteria led to representation of may be occurred between economical and technicaloperational
the modied version of Nicholas technique by him. In his modied criteria.
technique, three groups of weighting factors are used for ore The mining method which is not suitable from a technical
geometry, ore zone, hanging wall and foot wall. There were yet operational viewpoint denitely will not be protable. In other
essential problems with Nicholas technique [44]. These problems words, a mining engineer will not investigate a mining method
will be cleared in Section 5 by resolving a numerical problem. from an economical viewpoint unless he has found it suitable from
In Nicholas technique, mining methods are prioritized. The rst a technicaloperational point of view. For this reason, inspired by
three of them are considered suitable from a technicaloperational Nicholas technique, HEM is considered separately.
viewpoint. Since the applied criteria in Nicholas technique are of
technical type, the outcome of his technique will be mining methods 5. Case study: Western ore body of North anomaly of
which are prioritized based on technical criteria. But it is worth Choghart iron mine
noting that the costs of these methods are different from each other
in practice. After prioritizing, the rst three mining methods must In order to compare the results of the proposed approach with
also be investigated from an economical viewpoint. Finally, the most those of Nicholas technique, and investigate them, Western ore
economical mining method is selected as the most applicable and body of North anomaly of Choghart iron mine for which no mining
protable [44]. As a result unlike the previously introduced method has already been selected, is considered as a case study.
approaches, Nicholas technique is able to introduce just a unique This resource is located in 10 km at North West of Choghart mine in
mining method as the most practical and economical. A summary of central province of Yazd, Iran [30]. The model of this mineral
MMS process using Nicholas technique is represented in Tables 37. resource and its situation related to ground level is shown in Fig. 3.
1054 A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061

Table 36
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D1c22 ).

D1c22 BC SLC C&F OP SLS


1 1 1
BC (1/1, 1/1) (1/1, 2/3) (5/6, 6/7) (5/6, 6/7) (1/1, 2/3)1
SLC (1/1, 2/3) (1/1, 1/1) (2/3, 4/5)1 (3/4, 5/6)1 (1/1, 2/3)
C&F (5/6, 6/7) (2/3, 4/5) (1/1, 1/1) (1/1, 2/2)1 (2/3, 4/5)
OP (5/6, 6/7) (3/4, 5/6) (1/1, 2/2) (1/1, 1/1) (3/4, 5/6)
SLS (1/1, 2/3) (1/1, 2/3)1 (2/3, 4/5)1 (3/4, 5/6)1 (1/1, 1/1)

Table 37
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D2c22 ).

D2c22 BC SLC C&F OP SLS

BC (1/1, 1/1) (2/3, 3/4)1 (5/6, 6/6)1 (5/6, 6/7)1 (1/1, 2/2)1
SLC (2/3, 3/4) (1/1, 1/1) (1/2, 2/3)1 (2/2, 3/4)1 (1/2, 3/3)
C&F (5/6, 6/6) (1/2, 2/3) (1/1, 1/1) (1/2, 3/3)1 (6/6, 6/7)
OP (5/6, 6/7) (2/2, 3/4) (1/1, 2/3) (1/1, 1/1) (6/6, 7/7)
SLS (1/1, 2/2) (1/2, 3/3)1 (6/6, 6/7)1 (6/6, 7/7)1 (1/1, 1/1)

Table 38
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D3c22 ).

D3c22 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/2, 2/3)1 (5/6, 6//7)1 (3/4, 5/6)1 (2/2, 3/4)1
SLC (1/2, 2/3) (1/1, 1/1) (2/2, 3/4)1 (1/2, 2/2)1 (1/1, 2/2)1
C&F (5/6, 6//7) (2/3, 4/4) (1/1, 1/1) (1/2, 2/3) (2/2, 2/3)
OP (3/4, 5/6) (1/2, 2/2) (1/2, 2/3)1 (1/1, 1/1) (1/1, 2/3)
SLS (2/2, 3/4) (1/1, 2/2) (2/2, 2/3)1 (1/1, 2/3)1 (1/1, 1/1)

Table 39
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D4c22 ).

D4c22 BC SLC C&F OP SLS

BC (1/1, 1/1) (2/2, 3/4)1 (5/6, 6/7)1 (6/6, 6/7)1 (2/3, 4/5)1
SLC (2/2, 3/4) (1/1, 1/1) (1/2, 2/3)1 (2/3, 4/5)1 (1/1, 2/2)
C&F (5/6, 6/7) (1/2, 3/4) (1/1, 1/1) (1/1, 2/3)1 (1/2, 3/3)
OP (6/6, 6/7) (2/3, 4/5) (1/1, 2/3) (1/1, 1/1) (2/3, 4/5)
SLS (1/1, 2/3) (1/1, 2/2)1 (1/2, 3/3)1 (2/3, 4/5)1 (1/1, 1/1)

Table 40
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D5c22 ).

D5c22 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/2, 2/3)1 (3/4, 4/4)1 (5/6, 6/7)1 (1/1, 2/3)1
SLC (1/2, 2/3) (1/1, 1/1) (2/2, 2/3)1 (2/3, 4/5)1 (1/2, 2/3)
C&F (3/4, 4/4) (2/2, 2/3) (1/1, 1/1) (1/1, 2/3)1 (3/4, 4/5)
OP (5/6, 6/7) (2/3, 4/5) (1/1, 2/3) (1/1, 1/1) (3/4, 5/5)
SLS (2/3, 3/4) (1/2, 2/3)1 (3/4, 4/5)1 (3/4, 4/5)1 (1/1, 1/1)

Table 41
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D6c22 ).

D6c22 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/2, 2/3)1 (3/4, 4/4)1 (5/6, 6/7)1 (1/1, 2/3)1
SLC (1/2, 2/3) (1/1, 1/1) (2/2, 2/3)-1 (2/3, 4/5)1 (1/2, 2/3)
C&F (3/4, 4/4) (2/2, 2/3) (1/1, 1/1) (1/1, 2/3)1 (3/4, 4/5)
OP (5/6, 6/7) (2/3, 4/5) (1/1, 2/3) (1/1, 1/1) (3/4, 5/5)
SLS (2/3, 3/4) (1/2, 2/3)1 (3/4, 4/5)1 (3/4, 4/5)1 (1/1, 1/1)

Its other features are represented in Table 10. As shown in Fig. 3, thickness is steady. Its slope varies from 458 to 808. Its thickness also
the resource has three mineralized zones which are nominated by varies from 20 m to 60 m. In this regard, geometrical and also
ore bodies 1, 2 and 3. Many problems resulted from uncertainty and geological characteristics of all mineral resources around the world
ambiguity in geometrical characteristics of all three ore bodies can are nearly similar to each other. Assigning a certain crisp number to
be observed. For example, as shown in Table 10 and Fig. 3, for ore slope, thickness, etc. of a mineral resource is therefore a futile task.
body No. 1, no steady slope can be observed. Neither its slope nor its This indicates the ambiguous and unsteady nature of mineral
A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061 1055

Table 42
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D7c22 ).

D7c22 BC SLC C&F OP SLS


1 1
BC (1/1, 1/1) (1/1, 2/2) (3/3, 4/5) (1/1, 2/3) (1/1, 2/2)1
SLC (1/1, 2/2) (1/1, 1/1) (3/3, 4/5) (1/1, 2/2)1 (1/1, 1/2)1
C&F (3/3, 4/5)1 (3/3, 4/5)1 (1/1, 1/1) (3/4, 4/5)1 (3/4, 4/4)1
OP (1/1, 2/3) (1/1, 2/2) (3/4, 4/5) (1/1, 1/1) (1/1, 2/2)
SLS (1/1, 2/2) (1/1, 1/2) (3/4, 4/4) (1/1, 2/2)1 (1/1, 1/1)

Table 43
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D1c23 ).

D1c23 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/2, 2/3)1 (5/5, 6/6)1 (1/1, 2/2) (1/2, 2/2)1
SLC (1/2, 2/3) (1/1, 1/1) (3/3, 4/5)1 (1/2, 3/4) (1/1, 1/2)1
C&F (5/5, 6/6) (3/3, 4/5) (1/1, 1/1) (5/6, 6/7) (2/2, 3/4)
OP (1/1, 2/2)1 (1/2, 3/4)1 (5/6, 6/7)1 (1/1, 1/1) (1/2, 3/3)1
SLS (1/2, 2/2) (1/1, 1/2) (2/2, 3/4)1 (1/2, 3/3) (1/1, 1/1)

Table 44
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D2c23 ).

D2c23 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/2, 3/3)1 (4/5, 6/7)1 (5/6, 6/6)1 (1/2, 3/4)1
SLC (1/2, 3/3) (1/1, 1/1) (2/2, 3/4)1 (2/3, 3/4)1 (1/1, 1/2)
C&F (4/5, 6/7) (2/2, 3/4) (1/1, 1/1) (1/1, 2/3)1 (1/2, 2/3)
OP (5/6, 6/6) (2/3, 3/4) (1/1, 2/3) (1/1, 1/1) (2/3, 3/4)
SLS (1/2, 3/4) (1/1, 1/2)1 (1/2, 2/3)1 (2/3, 3/4)1 (1/1, 1/1)

Table 45
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. D3c23 ).

D3c23 BC SLC C&F OP SLS

BC (1/1, 1/1) (1/2, 2/2)1 (3/4, 4/5)1 (7/8, 8/8)1 (1/2, 2/2)1
SLC (1/2, 2/2) (1/1, 1/1) (1/2, 2/3)1 (3/4, 4/5)1 (1/1, 1/2)
C&F (3/4, 4/5) (1/2, 2/3) (1/1, 1/1) (1/2, 2/3)1 (1/2, 2/3)
OP (7/8, 8/8) (3/4, 4/5) (1/2, 2/3) (1/1, 1/1) (3/4, 4/5)
SLS (1/2, 2/2) (1/1, 1/2)1 (1/2, 2/3)1 (3/4, 4/5)1 (1/1, 1/1)

Table 46
Final weights of HTOM alternatives in form of fuzzy and crisp values. Three more suitable methods which are attained more weight has been considered for HEM.

Alternatives Final fuzzy weights based a level cut0.5 Crisp results (X*)
on technical and operational
conditions w a < jL < b g < jR < d
BC H (0.04/0.12, 0.37/1.08) 0.08 0.66 0.44
SLC (0.02/0.06, 0.19/0.61) 0.04 0.36 0.24
C&F H (0.02/0.06, 0.22/0.66) 0.04 0.39 0.26
OP H (0.09/0.26, 0.79/2.14) 0.16 1.35 0.88
SLS (0.01/0.05, 0.16/0.48) 0.03 0.29 0.19

Table 47 Table 48
Fuzzy pair-wise comparison matrix of criteria (i.t.o. A). Fuzzy pair-wise comparison matrix of sub-criteria (i.t.o. B1).

A B1 B2 B3 B1 C11 C12 C13

B1 (1/1, 1/1) (2/3, 4/5) (1/1, 2/3) C11 (1/1, 1/1) (1/1, 2/3)1 (5/6, 7/8)
B2 (2/3, 4/5)1 (1/1, 1/1) (1/2, 3/4)1 C12 (1/1, 2/3) (1/1, 1/1) (7/8, 9/9)
B3 (1/1, 2/3)1 (1/2, 3/4) (1/1, 1/1) C13 (5/6, 7/8)1 (7/8, 9/9)1 (1/1, 1/1)

resources. It is obvious that it is much more accurate to specify these Mining Industries Development and Renovation Organization
characteristics by fuzzy numbers than by crisp numbers. (IMIDRO) which all iron mines in Iran are supported by it. This
Mining companies know the importance of MMS problem to paper also has been carried out by support of IMIDRO Co. to nd
start up the mining operations. They use the approaches which are out a solution to start the mining operations in North anomaly of
introduced to modify the MMS problem, because it would be so Choghart iron mine in Iran. No mining method is applied for this
protable for a mining company if its mining projects be started by anomaly yet, and result of proposed approach in this study is
a perfect MMS procedure. One of them is the Iranian Mines and applicable for IMIDRO Co.
1056 A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061

Table 49 Table 51
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. B2). Fuzzy pair-wise comparison matrix of alternatives (i.t.o. C12).

B2 BC C&F OP C12 BC C&F OP


1 1
BC (1/1, 1/1) (2/2, 3/4) (7/7, 8/9) BC (1/1, 1/1) (1/2, 2/3) (1/1, 2/2)1
SLC (2/2, 3/4) (1/1, 1/1) (3/3, 4/5)1 C&F (1/2, 2/3)1 (1/1, 1/1) (2/3, 3/4)1
OP (7/7, 8/9) (3/3, 4/5) (1/1, 1/1) OP (1/1, 2/2) (2/3, 3/4) (1/1, 1/1)

Table 50 Table 52
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. C11). Fuzzy pair-wise comparison matrix of alternatives (i.t.o. C13).

C11 BC C&F OP C13 BC C&F OP


1 1
BC (1/1, 1/1) (1/2, 2/3) (1/2, 2/3) BC (1/1, 1/1) (3/4, 4/5) (1/1, 2/2)1
C&F (1/2, 2/3) (1/1, 1/1) (3/4, 4/5) C&F (3/4, 4/5) (1/1, 1/1) (1/2, 2/3)
OP (1/2, 2/3)1 (3/4, 4/5)1 (1/1, 1/1) OP (1/1, 2/2) (1/2, 2/3)1 (1/1, 1/1)

5.1. Resolving the problem by Nicholas technique deffuzication of nal fuzzy weight for BC mining alternative
based on Eq. (20). The nal crisp weights of other mining
As mentioned earlier, in Nicholas technique, a weight is alternatives have been calculated similarly.
assigned to each mining method considering the scores used for R 0:12 x0:04 R 0:37 R 1:08 1:08x
 0:04 0:08 xdx 0:12 1xdx 0:37 0:71 xdx
different condition of mineral resources, and by adding up these XTOHM BC
R 0:12 x0:04 R 0:37 R 1:08 1:08x
weights, the ultimate weight of that mining method is obtained. 0:04 0:08 dx 0:32 1dx 0:37 0:71 dx

These scores are shown in Tables 46. For example, the average 0:44
thickness of north anomaly, as shown in Table 10, is 40 m. So, as
shown in Table 4, the score of 2 is allocated to block caving (BC) Graph of nal fuzzy weight for mining alternatives and the bar
mining method. Because the thickness of 40 m is located between graph indicating the amount of inuence of each of the sub-criteria
30 m to 100 m in Table 4 of Nicholas. With this explanation, other on the nal weight of each mining method are shown in Figs. 5 and
mining methods are similarly weighted. The obtained weights are 6 respectively.
summarized in Table 11 for ve different mining methods. Considering Table 46, from a technicaloperational view point,
By adding up the related scores, the ultimate weights of 33, 33, mining alternatives of OP, BC and C&F has obtained the three most
28, 46 and 27 are obtained for mining methods of block caving (BC), weights respectively. Using HEM, these three alternatives must
sub-level caving (SLC), cut and ll (C&F), open pit (OP) and sub- also be compared to each other from an economical view point.
level stoping (SLS), respectively. The results of Nicholas technique For this purpose, as the second hierarchical model, HEM and
indicate that OP is the most suitable mining method for three mentioned mining alternatives are shown in Fig. 7.
exploitation of North anomaly of Choghart mine. Like HTOM, fuzzy comparison matrices of HEM are represented
in Tables 4755.
5.2. Resolving the problem using the proposed approach
5.2.2. Results of HEM
As shown in Figs. 4 and 7, based on the concept of hierarchical Local fuzzy and crisp weights of criteria and mining alternatives
analysis shown in Tables 8 and 9, the hierarchical models of HEM have been calculated like HTOM. The nal fuzzy and crisp weights
and HTOM are constructed for resolving MMS problem in North of mining alternatives are also represented in Table 56. As shown in
anomaly of Choghart. Like Nicholas technique, here ve mining Table 56, OP is found as the most protable mining alternative
methods including BC, SLC, C&F, OP and SLS are considered as which is the most suitable alternative for mining of north anomaly
alternatives of HTOM model, so that three mining methods with of Choghart mine.
the most weight, as the alternatives of HEM model, are further Graph of nal fuzzy weight for mining alternatives and the
investigated economically. inuence of each sub-criteria of HEM on the nal weight of mining
Now, considering HTOM model, the fuzzy comparison matrices alternatives are also shown in Figs. 8 and 9.
must be constructed based on the engineering judgments of DM The process of selecting the most suitable mining method using
and according to Bukleys FAHP method. These judgments are proposed approach is summarized in Fig. 10 schematically. This
represented in the form of fuzzy comparison matrices in Tables gure indicates the introduction of OP for starting mining of North
1245. anomaly of Choghart mine. It is the most suitable mining method,
Using a computer program, the local fuzzy weights of each technicaloperationally as well as economically.
criterion, sub-criterion and mining alternative of HTOM model is
obtained. Their local crisp weights are also obtained using Eq. (20). 5.2.3. Calculation of the inconsistency ratio of fuzzy comparison
matrices
5.2.1. Results of HTOM Consistency ratio of comparison matrices has been checked by
Up to now, the ve mentioned mining methods are prioritized conception of Theorem 1. According to Theorem 1 we know
using HTOM. As the multiplication of trapezoidal fuzzy numbers by bi j ; g i j   bik ; g ik :bk j ; g k j .
each other will not equal a trapezoidal fuzzy number, a level-cut of For example, considering Table 14 we have b13 ; g 13  
0.5 is used in order to depict the fuzzy nal weight graph. Value b12 ; g 12  b23 ; g 23  ) 6; 7  3; 4 2; 3 ) 6; 7  6; 12 for
corresponding to membership function of 0.5 on the right side and i; j; k 1; 3; 2 respectively. Then, it is perfectly consistent. Another
left side of each trapezoidal graph is shown by jL, jR respectively. example is about Table 49 of HEM where we have
The results of nal fuzzy and crisp weights and also the value of b32 ; g 32   b31 ; g 31  b12 ; g 12  ) 3; 4  7; 8 13;
1
jL, jR for each mining alternatives is represented in Table 46. 2 ) 3; 4  2:33; 4. It is also perfectly consistent for i; j; k 3; 2; 1.
As an example, the way of calculation of nal crisp weight for BC In Table 36, for i; j; k 1; 5; 4, we have 12; 1  4; 5 16; 16 )
mining alternative is as follows. This crisp weight is obtained from 0:5; 1 6 0:66; 0:83 which is not consistent perfectly but it
A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061 1057

Table 53 not a certain geometrical shape so that part of it may be nominated


Fuzzy pair-wise comparison matrix of alternatives (i.t.o. C14).
by state M and another part by state T/P. Therefore, DM would
C14 BC C&F OP encounter the dilemma in using one of 2 or 3 values for OP method.
BC (1/1, 1/1) (2/3, 3/4) (1/2, 2/3)1 Another reason indicating the inability of Nicholas technique in
C&F (2/3, 3/4)1 (1/1, 1/1) (5/6, 6/6)1 selecting the suitable mining alternative is that a change in the
OP (1/2, 2/3) (5/6, 6/6) (1/1, 1/1) state of criteria accounts for a huge difference between their
scores. As an example, as shown in Table 3, for C&F mining method,
two values of 4 and 0 are considered as scores assigned to sub-
Table 54 criteria of Intermediate and Thick as parts of thickness
Fuzzy pair-wise comparison matrix of alternatives (i.t.o. C15).
criterion respectively. Considering the footnote of Table 3, if the
C15 BC C&F OP thickness of resource changes from 30 m to 31 m, its score will
BC (1/1, 1/1) (1/2, 2/3) (7/8, 8/8)1 change from the best state of +4 to the worst state of 0. That is, a
C&F (1/2, 2/3)1 (1/1, 1/1) (4/4, 4/5)1 1 m change in thickness, accounts for a big change in score. As a
OP (7/8, 8/8) (4/4, 4/5) (1/1, 1/1) result, not only the problem of putting a border between two sub-
criteria of thick and intermediate exist, but also the assign-
ment of scores to these categories has been carried out incorrectly.
Table 55 In addition, the two sub-criteria of thick and intermediate,
Fuzzy pair-wise comparison matrix of sub-criteria (i.t.o. B3).
share the value of 30. In other words, either scores of 0 or 4 may be
B3 C14 C15 allocated to a resource with a thickness of 30 m. Another point
C14 (1/1, 1/1) (7/8, 9/9) worth mentioning is that in Nicholas technique allocation of
C15 (7/8, 9/9)1 (1/1, 1/1) judgments has been carried out using crisp numbers. These
numbers are not able to indicate the complicated and important
conditions of mineral resources properly. Another important point
is consistent reasonably. More descriptions are available in is that for DM, allocation of his own judgments is not optional.
[11,12,19]. Because in Nicholas technique, these scores are predened and DM
is just able to adapt the scores of the tables of Nicholas to the
5.3. Comparison of the results of Nicholas technique and conditions of the supposed mineral resource. On the other hand,
proposed approach the range of those scores is small and limited. It means that the
difference between favorable and unfavorable scores is little.
Resolving MMS problem by the use of both Nicholas technique Because the scores of 0, 4 are assigned to a resource in the worst
and proposed approach for the same mineral resource makes the and best state respectively. The lack of some of the important and
comparison of these two methods simpler. Although both methods inuential criteria in Nicholas technique would be another reason.
introduce OP as the most suitable alternative, there are many Table 57 indicates the results of resolving MMS problem in North
reasons indicating that Nicholas technique is not able to select the anomaly of Choghart mine by using both of the mentioned method.
most suitable alternative correctly. One of these reasons is the way Since all of the criteria of Nicholas technique are technical
of dening criteria and sub-criteria in Nicholas technique. operational, its results must be compared to those of HTOM. As
Generally, considering the uncertainties of mineral resources shown in Table 57, in Nicholas technique, two alternatives of BC
mentioned before, classication of criteria into two or three and SLC have obtained the same weight of 33. It means that
categories would not be correct. Because geometrical and Nicholas technique cannot prioritize mining alternatives properly.
geological condition of mineral resources is such ambiguous that In addition to aforementioned reasons, the most important reason
it is impossible to categorize them. For example, in Table 3, three of this deciency would be the use of an improper weighting
sub-criteria of M, T/P and I are taken for general shape criterion of system and the lack of determination of a weight for each criterion.
a mineral resource. As described earlier, a mineral resource has Even in the modied version of Nicholas, in spite of using

Table 56
Final fuzzy and crisp weights of three suitable mining method using HEM to select the optimum MMS. OP mining method is introduced as optimum mining method for
western part of north anomaly of Choghart iron mine.

Alternatives Final fuzzy weights based a level cut Crisp results (X*)
on economical conditions w
a < jL < b g < jR < d
BC (0.05/0.15, 0.49/1.36) 0.09 0.86 0.55
C&F (0.05/0.13, 0.43/1.23) 0.08 0.76 0.50
OP H (0.12/0.28, 0.84/1.99) 0.19 1.35 0.85

Table 57
Results Nicholas technique and proposed approach.

Nicholas MMS technique Proposed MMS technique

Mining methods Final weight Ranking results Mining Final weight Ranking results
methods
TOHM EHM TOHM EHM Total

OP 46 1 OP 0.88 0.85 1 1 1
BC 33 2 BC 0.44 0.55 2 2 2
SLC 33 2 C&F 0.26 0.50 3 3 3
C&F 28 3 SLC 0.24 4
SLS 27 4 SLS 0.19 5
1058 A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061

weighting factor, this deciency has not been completely

equations. It has a huge computational

Extent analysis method may be cannot


It is capable to apply triangular fuzzy

It is capable to apply triangular fuzzy


The method is dened for triangular
Its computational procedure is huge
Its computational procedure is high
eliminated. Because all of the criteria assume constant weights

procedure even for small problems

estimate always true weights [58]


It was not a solution to the linear
while the importance of mining alternatives are different for
various resources. This problem may be investigated another way.
In 1980s, when Nicholas introduced his technique, some of the
mining methods like SLC were not as much prevalent as today.
Therefore the way of scoring in these mining methods in Nicholas

numbers merely

numbers merely
technique were based on the amount of their importance and

fuzzy number
applicability, regarding the conditions of that time. But scoring

Deciencies
should not be dependent on time. It must be usable by DM at any
time. All of these reasons indicate that DM cannot rely on the
results obtained from Nicholas technique. In the proposed
approach, by using FAHP, all of the mentioned problems have
been nearly eliminated, so that it is possible for DM to allocate his

More exibility for DMs to express their

It is able to model multiple DMs judges


judgments to a mineral resource, considering the conditions of the

judges by trapezoidal fuzzy numbers

It can revise Juang and Lees fuzzy


time, such as the amount of applicability of mining methods, the

AHP has been developed to fuzzy

Its computational procedures is


price of mineral materials, etc.

environment by their method

It does not involve additional


Ease of its extension in fuzzy
6. Discussion

number 9 accurately
All of the resources having an outcrop, or having a negligible

relatively low
amount of overburden, the weight of sub-criteria of depth for some

environment
Advantages

operations
of the mining alternatives increase greatly. OP is one of these
alternatives. The existence of the outcrop on the ground level
suggests a depth of zero for the resources. The importance of this
matter for OP method is that the depth of zero causes a great
decrease in the stripping ratio. As a result, the cost of stripping and
development of the mine for OP alternative will be much less, in

It was introduced to drive a simple and

respect to each element under a certain


Representation the weight vectors with
the cases of missing data and multiple
Presentation of a method to produce a
unique fuzzy weight and applicable to
comparison to other mining alternatives. This decrease of cost has
a direct inuence on criterion B2 in Table 9. So, the return of initial
signicant factors via pair-wise
It was developed to weight the

aggregation of alternatives and


Determination of nal score of
alternatives by an appropriate
investment will be much quicker. Without stripping costs, a short

general algorithm for FAHP


time after its start up, the mine will reach to production and
income. This would have also an inuence on some of the sub-
criteria represented in Table 8. In other words, the depth of zero

attributes weights
leads to an increase of the weight of sub-criteria of D4c22 , D3c21 and
D1c21 . Because of the exceptional conditions posed by the existence
comparison

estimates

of outcrop for the resource, in the Op mining alternative, this

criterion
alternative in most criteria and sub-criteria will obtain a greater
Issue

weight in comparison to other mining alternatives. Because of this


great difference between OP and other mining methods, both
proposed approach and Nicholas technique have introduced OP as

Proposed method is based on triangular


It was a fuzzy version of saatys method
extended by De Gran and Lootsma as

the most suitable mining alternative.

analysis method for synthetic extent


It was proposed based on geometric

Laarhoven and Pedrycz method by

The method was used the interval


using minimization of logarithmic
It was a modifying version of Van

It was introduced by using extent


But the shortcoming of Nicholas technique in selecting the most
Use of trapezoidal fuzzy numbers
A brief history of proposed approaches on the FAHP and its principle problems.

arithmetic and a cut to estimate


Use of triangular fuzzy numbers

values of comparison matrices


suitable alternative will be obvious when the conditions of
mean method of Saatys AHP

resource have not created such a great difference among other


mining alternatives. In such a case, the weights of mining
eigenvector of matrices

alternatives would be close to each other.


regression function
rst FAHP method

Because of deciencies of its weighting system, Nicholas


fuzzy numbers
Characteristics

technique is not able to introduce the best alternative. As a result,


as discussed in previous chapter, problems such as obtaining equal
weights for two mining alternatives will occur.

7. Conclusions

This paper, as a modied version of Nicholas technique,


1983

1985

1989

1994

1996
Year

represents an approach to resolve the MMS problem. In proposed


approach, by the use of AHP and the control of inconsistency ratio
of judgments, weighting process has been modied in comparison
to Nicholas technique and problem such as obtaining equal nal
weights for each alternative has therefore been eliminated. For
Van Laarhoven and

Boender et al. [8]

DM, the allocation of engineering judgments has rather more


Pedrycz [56]

compatibility with the characteristics of mineral resources.


Buckley [12]

Cheng [16]

Chang [15]

Considering the uncertainty of characteristics of mineral resources,


Author(s)

and because of use of fuzzy numbers, the proposed approach can


Table 58

be used instead of previous techniques, the results of which were


not such reliable for the DM. We urge strongly that, due to
Table 58 (Continued )

Author(s) Year Characteristics Issue Advantages Deciencies

Cheng [17] 1996 It was proposed based on the grade Representation of new algorithm which It was presented to improve AHP and to Entropy is used when probability
value of membership function using is used the experience of expert to smooth evaluation process distribution is known [13]
triangular fuzzy numbers represent judgmental object.
Aggregate weights is calculated by It is exible, thoughtfulness and The method is based on both
entropy concepts efciency for DMs probability and possibility measures
[13]
It was introduced fuzzy standard which The computational requirement is not
is membership function of judgment tremendous

A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061


criteria for all sub items

Xu [61] 2000 It was applied last square priority It was proposed to Estimate the weights The numerical difculties can be It may not be to give vague judgments
method in fuzzy environment of factors from fuzzy judgment avoided in fuzzy matrices because it when comparing the relative
matrices correspond to the classical yields an analytic expression for the signicance of two factors
Last square priority method fuzzy weights
Csutora and Buckley [19] 2001 It was the fuzzication version of Less fuzziness of nal fuzzy weights is It is easy to use computationally Less fuzziness issue is not always
Saatys lmax method the methods issue applicable for all fuzzy weights
Can be handle for any type of fuzzy
numbers

Mikhailov [39] 2003 Fuzzy preference programming Deriving optimum crisp priorities from It does not need to construct reciprocal The mathematical complexity involved
method was developed by proposed fuzzy pair-wise comparison matrices may restrict its practicability [38]
approach
It can be applied for group decision Priorities can be derived from an It automatically guarantees the
making incomplete set of fuzzy judgments satisfaction of fuzzy condition of order
using a-cuts preferences at each level, if the interval
judgments with regards to each a-cut
level are consistent [5,14]
It can employ specic forms of fuzzy
sets and does not require a nal ranking
procedure to compare the fuzzy scores

Wang et al. [59]. 2006 It was modied fuzzy LLSM wich was Elimination drawbacks of conventional It can derive normalized triangular It is proposed to use by triangular fuzzy
proposed by Van Laarhoven and LLSM method of Van Laarhoven and fuzzy weights from complete or numbers
Pedrycz Pedrycz and Boender et al. incomplete comparison matrices.
LLSM was formulated as a constrained It is able to solve the complex MCDM
nonlinear optimization model in their problems
method

1059
1060 A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061

ambiguities and uncertainties in mineral resources, the proposed [24] C. Guray, C. Nese, V. Atalaya, H.E. Gunhan Pasam, A. Metoglu, Ore-age: a hybrid
system for assisting and teaching MMS, Expert Systems with Applications 24
approach have been represented based on fuzzy MCDM tools, (2003) 261271.
which are of suitable methods of decision making. [25] H. Hamrin, Choosing an Underground Mining Method. Underground Mining
Resolving the MMS problem using mathematical modeling or Methods Handbook, AIME, New York, 1982.
[26] H.L. Hartman, Introductory Mining Engineering, rst edition, John Wiley, New
evolutionary algorithms may be regarded as future studies. York, 1987.
[27] J.A. Hudson, J.P. Harrison, Engineering Rock Mechanics: An Introduction to the
Principles, fourth impression, Elsevier Publisher, 2005, pp. 163.
Acknowledgments [28] I. Jeffreys, The use of compensatory and non-compensatory multi-criteria analysis
for small-scale forestry, small-scale forest economics, Management and Policy 3
The authors are deeply thankful for supports of IMIDRO and (1) (2004) 99117.
[29] C. Kahraman, D. Ruan, Y. Dogan, Fuzzy group decision making for facility location
KAVOSHGARAN Co., Tehran, Iran. They also express their sincere selection, Information Sciences 157 (2003) 135153.
thanks to the referees for their helpful suggestions and recom- [30] Kavoshgaran Consulting Engineers, Feasibility study of north anomaly of Cho-
mendations. ghart iron ore-western ore body, 2005, pp. 193.
[31] A. Kesimal, A. Bascetin, Application of fuzzy multiple attribute decision making in
mining operations, Mineral Resources Engineering 11 (1) (2002) 5972 (Imperial
College Press).
[32] O. Kulak, C. Kahraman, Fuzzy multi-attribute selection among transportation
Appendix A companies using axiomatic design and analytic hierarchy process, Information
Sciences 170 (2005) 191210.
See Table 58. [33] D.H. Laubscher, Selection of Mass Underground Mining Methods, Design and
Operation of Caving and Sublevel Stoping Mines, D. Stewart, SME-AIME, New
York, 1981 (Chapter 3, pp. 2338).
References [34] M. Lee, H. Pham, X. Zhang, A methodology for priority setting with application to
software development process, European Journal of Operational Research 118
(1999) 375389.
[1] L. Adler, S.D. Thompson, Mining method classication systems, SME Mining [35] L.C. Leung, D. Cao, On consistency and ranking of alternatives in fuzzy AHP,
Engineering Handbook, Society for Mining Engineering, Metallurgy and Explora- European Journal of Operational Research 124 (2000) 102113.
tion, Inc., 1992, pp. 531537. [36] W.-X. Li, Applications of Fuzzy Mathematics in Mining and Geotechnical
[2] A. Almeida, L.H. Alencar, C.M.G. Miranda, Mining methods selection based on Engineering, The Press of Metallurgy Industry, Beijing, 1998.
multi criteria models. Mining methods selection based on multicriteria models, [37] I. Linkov, A. Ramadan, Comparative Risk Assessment and Environmental Decision
in: Proceedings of the 32nd International Symposium on Computer Applications Making, Kluwer Publisher, 2004, pp. 15-54.
in the Minerals Industry, Taylor & Francis Group, Arizona, 2005. [38] D. Metin, I. Yuksel, Developing a fuzzy analytic hierarchy process (AHP) model for
[3] S. Alpay, M. Yavuz, A Decision Support System for Underground Mining Method behavior-based safety management, Information Sciences 178 (2008) 1717
Selection, Lecture Notes in Computer Science, Springer, Berlin, Heidelberg, 2007, 1733.
pp. 334343. [39] L. Mikhailov, Deriving priorities from fuzzy pair wise comparison judgments,
[4] K.M. AlSubhi AlHarbi, Application of the AHP in project management, Interna- Fuzzy Sets and Systems 134 (2003) 365385.
tional Journal of Project Management 19 (2001) 1927. [40] L. Miller-Tait, R. Pakalnis, R. Poulin, UBC mining method selection, Mine Planning
[5] C.A. Banae Costa, J.C. Vansnick, A critical analysis of the Eigen value method used and Equipment Selection (1995) 163168.
to derive priorities in AHP, European Journal of Operational Research 187 (2008) [41] R.G.K. Morrison, A Philosophy of Ground Control, McGill University, Montreal,
14221428. Canada, 1976, pp. 125159.
[6] T.C. Bibb, K.M. Hargrove, Coal Mining: Method Selection, SME Mining Engineering [42] M.Z. Naghadehi, R. Mikaeil, M. Ataei, The application of fuzzy analytic hierarchy
Handbook, second edition, Society for Mining Engineering, Metallurgy and process (FAHP) approach to selection of optimum underground mining method
Exploration, Inc., 1992, pp. 18551866. for Jajarm Bauxite Mine, Iran, Expert Systems with Applications (2008),
[7] M.R. Bitarafan, M. Ataei, Mining method selection by multiple criteria decision doi:10.1016/j.eswa.2008.10.006.
making tools, The Journal of The South African Institute of Mining and Metallurgy [43] D.E. Nicholas, Selection ProcedureA Numerical Approach, Design and Operation
(2004) 493498. of Caving and Sublevel Stoping Mines, Society of Mining Engineers of the Amer-
[8] C.G.E. Boender, J.G. De Graan, F.A. Lootsma, Multi criteria decision analysis with ican Institute of Mining, Metallurgical, and Petroleum Engineers, Inc. Port City
fuzzy pair wise comparison, Fuzzy Sets and Systems 29 (1989) 133143. Press, 1981, pp. 3953.
[9] S.H. Boshkov, F. Wright, Basic parametric criteria in the selection, design and [44] D.E. Nicholas, Selection Procedure, SME Mining Engineering Handbook, second
development of underground mining systems, in: A.B. Cummins, I.A. Given (Eds.), edition, Society for Mining Engineering, Metallurgy and Exploration, Inc., 1992,
SME Mining Engineering Handbook, vol. 1, SME-AIME, New York, 1973, pp. 12.2 pp. 20902106.
12.13. [45] T.A. Ohara, S.C. Suboleski, Costs and cost estimation, SME Mining Engineering
[10] B.H.G. Brady, E.T. Brown, Rock Mechanics for Underground Mining, George Allen Handbook, second edition, Society for Mining Engineering, Metallurgy and
and Unwin, London, 1985. Exploration, Inc., 1992, pp. 405424.
[11] J.J. Buckley, T. Feuring, Y. Hayashi, Fuzzy hierarchical analysis revisited, European [46] M. Osanloo, M. Ataei, M. Heidari, Selection of mining method for anomaly No 3 of
Journal of Operational Research 129 (2001) 4864. Gol-Gohar iron mine of Iran, Mine Planning and Equipment Selection (2003) 105
[12] J.J. Buckley, Fuzzy hierarchical analysis, Fuzzy Sets and Systems 17 (1985) 233 108.
247. [47] S. Planeta, J. Paraszczak, J. Szymanski, Selection criteria for underground stoping
[13] G. Buyukozkan, C. Kahraman, D. Ruan, A fuzzy multi-criteria decision approach methods in narrow vein deposits, Mine Planning and Equipment Selection (2001)
for software development strategy selection, International Journal of General 233241.
Systems 33 (AprilJune (23)) (2004) 259280. [48] J.C. Pomerol, S. Barba Romero, Multi Criterion Decision in Management: Principles
[14] O. Cakir, On the order of the preference intensities in fuzzy AHP, Computers and and Practice, rst edition, Kluwer Academic Publishers, Norwell, 2000.
Industrial Engineering 54 (2008) 9931005. [49] T.L. Saaty, How to make a decision: analytic hierarchy process, European Journal
[15] C.H. Cheng, D.L. Mon, Evaluation weapon system by AHP based on fuzzy scales, of Operational Research 48 (1990) 926.
Fuzzy Sets and Systems 63 (1994) 110. [50] T.L. Saaty, The Analytic Hierarchy Process: Planning, Priority Setting, Resource
[16] C.H. Cheng, Evaluating naval tactical missile systems by fuzzy AHP based on the Allocation, McGraw-Hill, New York, 1980.
grade value of membership function, European Journal of Operational Research [51] J.B. Sheu, A hybrid fuzzy-based approach for identifying global logistics strategies,
96 (1996) 343350. Transportation Research Part E: Logistics and Transportation Review 40 (2004)
[17] D.Y. Chang, Applications of the extent analysis method on fuzzy AHP, European 3961.
Journal of Operational Research 95 (1996) 649655. [52] W. Siler, J. Buckley, Fuzzy Expert Systems and Fuzzy Reasoning, John Wiley &
[18] C. Ching-Hsue, Evaluating naval tactical missile systems by fuzzy AHP based on Sons, Inc., 2005, pp. 2954.
the grade value of membership function, European Journal of Operational [53] D.L. Surulescu, G. Chiril, S. Irimie, Theoretical Consideration in Establishing the
Research 96 (1997) 343350. Technical Parameters for Sublevel Caving, MPES, Trino, Italy, 2006.
[19] R. Csutora, J.J. Buckley, Fuzzy hierarchical analysis: the Lambda-Max method, [54] E. Tolga, M.L. Demircan, Operating system selection using fuzzy replacement
Fuzzy Sets and Systems 120 (2001) 181195. analysis and analytic hierarchy process, International Journal of Production
[20] H. Deng, Multicriteria analysis with fuzzy pair wise comparison, International Economics 97 (2005) 89117.
Journal of Approximate Reasoning 21 (1999) 215231. [55] F. Tunc, A.B. Bozbura, Prioritization of organizational capital measurement indi-
[21] T.S. Felix, N. Kumar, Global supplier development considering risk factors using cators using fuzzy AHP, International Journal of Approximate Reasoning 44
fuzzy extended AHP-based approach, Omega 35 (2007) 417431. (2007) 124147.
[22] M.A. Grima, Neuro-fuzzy Modeling in Engineering Geology, A.A. Balkema, Rot- [56] P.J.M. Van Laarhoven, W. Pedrycz, A fuzzy extension of Saatys priority theory,
terdam, The Netherlands, 2000, pp. 244. Fuzzy Sets and Systems 11 (1983) 229241.
[23] I. Gulfem, G. Buyukozkan, Using a multi-criteria decision making approach to [57] Y.-M. Wang, Taha M.S. Elhag, On the normalization of interval and fuzzy weights,
evaluate mobile phone alternatives, Computer Standards and Interfaces 29 (2007) Fuzzy Sets and Systems 157 (2006) 24562471.
265274.
A. Azadeh et al. / Applied Soft Computing 10 (2010) 10401061 1061

[58] Y.-M. Wang, Y. Luo, Z. Hua, On the extent analysis method for fuzzy AHP [61] R. Xu, Fuzzy least-squares priority method in the analytic hierarchy process,
and its applications, European Journal of Operational Research 186 (2008) Fuzzy Sets and Systems 112 (2000) 395404.
735747. [62] L. Zadeh, Fuzzy sets, Information Control 8 (1965) 338353.
[59] Y.-M. Wang, B. Taha, Z. Hua, A modied fuzzy logarithmic least squares method [63] K.-J. Zhu, Y. Jing, D.-Y. Chang, A discussion on extent analysis method and
for fuzzy analytic hierarchy process, Fuzzy Sets and Systems 157 (2006) applications of fuzzy AHP, European Journal of Operational Research 116
30553071. (1999) 450456.
[60] L. Xinchun, Z. Youdi, Mineral Resource Evaluation based on AHP, Mine Planning [64] H.-J. Zimmermann, Fuzzy Set Theory and its Applications, third edition, Kluwer
and Equipment Selection, Taylor & Francis Group, 2004, pp. 8588. Academic Publishers, 1996.