Resources Policy 69 (2020) 101879

Contents lists available at ScienceDirect

Resources Policy
journal homepage: http://www.elsevier.com/locate/resourpol

Technology adoption in mining: A multi-criteria method to select emerging

technology in surface mines
Oluwatobi Dayo-Olupona a, Bekir Genc a, *, Moshood Onifade a, b
a
The School of Mining Engineering, University of the Witwatersrand, Johannesburg, South Africa
b
Department of Mining and Metallurgical Engineering, University of Namibia, Ongwediva, Namibia

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

Keywords: The integration of technologies across the mining value chain is becoming critical, and it is crucial to identify the
AHP best possible technologies applicable to an operation, based on the chosen criteria. This study investigates the
Decision-making adoptable technologies for a mining project, develops a conceptual framework for the analytical process and
Fuzzy multiple attribute
validates the framework using a hypothetical case study. The case study comprised of a technology decision
MCDM
Mining technology
problem, the result of which consisted of six technology alternatives, four criteria, and one decision maker.
PROMETHEE Of the several Multi-Criterial Decision Making (MCDM) techniques available, the fusion of the analytic hi­
erarchy process (AHP) and preference ranking organisation method of enrichment (PROMETHEE) techniques
were used for this study. The AHP was used to determine the hierarchal weight of each decision-making criterion
and its consistency, while the PROMETHEE method was used to carry out the overall process evaluation.
Additionally, the fuzzy set theory was infused into the hierarchical structure analysis to evaluate the quantitative
economic criterion to curb uncertainty and imprecision. The results of the analysis show that the technology
alternative A3 – Artificial Intelligence (AI) – is the most preferred alternative, which outranks other comparative
technologies. The approach proposed in this study can help provide the basis for any technology adopting mining
company to build its technology business case, strategy and or roadmap to achieve the desired outcome.

1. Introduction/background mining approaches to an era where new technologies are being adopted
from other industries and the engineering discipline (Jang and Topal,
Mining is dynamic in nature and this has resulted in it being classi­ 2020). This transition, however, requires the emergence of new opera­
fied as a complex industry (Runge, 1995). Apart from its dynamism, tional models (Farrelly et al., 2012). Mining companies therefore need to
several other factors contribute to the complexity of the industry. Some recognize technology as a process enabler, and strategically prioritise
of them include fluctuating commodity prices; rising cost of operations; technological integration across the mining value chain (Deloitte,
continuing extraction of low ore grades; resource scarcity; resource 2019). Additionally, these technologies have the capacity to generate a
nationalism; increased environmental regulation; and geopolitical greater level of responsiveness, which will position mining companies
instability (Durrant-Whyte et al., 2015; World Economic Forum (WEF, for long-term growth (Deloitte, 2015). The proposition of Farrelly et al.
2017). (2012), Jang and Topal (2020), in line with Dickerson (1997) study
The cumulative effect of these identified challenges has resulted in a proposed a model to segment new technologies for easy adoption.
large-scale value destruction over the last 15 years (Bryant, 2015). Furthermore, based on a study of 41 mining projects, Readett (2006)
Resolving this value conundrum requires mining companies to consider outlined some principles for proper selection and implementation of
certain resolutions. Some of these resolutions include pushing the technology. Building on the aforementioned, Crawford (2018) show­
boundaries on digital transformation; attracting a truly diverse work­ cased 13 mine sites that are realising operational improvements, and
force; and re-envisioning corporate strategy. In this study, the focus was these value-additions were realized through the adoption of new and
on digital transformation aided by technology. emerging technologies. Even after these successful implementations, a
The mining industry is at the point of transition from conventional gap still exists; it is inherently unclear how each of the value adding

* Corresponding author. University of the Witwatersrand, Faculty of Engineering and the Built Environment, School of Mining Engineering, P.O. WITS, 2050,
Johannesburg, South Africa.
E-mail addresses: [email protected] (O. Dayo-Olupona), [email protected] (B. Genc), [email protected] (M. Onifade).

https://doi.org/10.1016/j.resourpol.2020.101879
Received 14 May 2020; Received in revised form 16 September 2020; Accepted 17 September 2020
0301-4207/© 2020 Elsevier Ltd. All rights reserved.
O. Dayo-Olupona et al. Resources Policy 69 (2020) 101879

technologies were selected, making replication of this success in other However, for this study, The MCDM tools are the most preferred
mines difficult. Amadi-Echendu et al. (2019) and Jordaan and Hendricks tools to assist in this situation. The tool is often used in operational
(2009) thus pointed out that, there is still a dire need for a coherent research because of its ability to scientifically solve problems that
technology roadmap and strategy to help facilitate technology adoption involve both quantitative and qualitative analysis (Sitorus et al., 2019).
and implementation within the mining industry. Filling this gap requires In addition, the scientific process allows the possibility to consider
an algorithmic process that will make the selection and deployment of complex and conflicting multiple uncertainties and criteria. Such sce­
technologies easy for the adopting mining company. narios would normally leave a decision maker with experience and
In keeping with the aforementioned, this study aimed to develop an intuition as a guide.
analytical algorithm that would help filter through an extensive list of This operation research method’s ability to solve critical operational
technologies with the aim of selecting the most suitable for a mining and non-hypothetical real-life problems within the mining industry and
operation. This analytical algorithm filtered through the well-organised other industries makes it an analytic tool of choice. MCDM is broadly
list of technologies reported by Jacobs (2016). This sorting was based on classified into MADM and MODM (Sabaei et al., 2015; Zavadskas et al.,
the criteria chosen by a mining company’s decision maker. Jacobs 2014; Hwang and Yoon, 1981). Of the several MCDM techniques, the
(2016) technology map consists of a comprehensive list of modern combination of AHP and PROMETHEE techniques have been used. The
technologies that can be adopted across the mining value chain with the justification for hybridizing these two MCDM techniques were based on
aim of creating value. Jacobs’ (2016) goal for creating the map was to each one’s peculiar characteristics. And these features were uniquely
enact a platform that provides technology information which specif­ used at different phases/stages of the computational evaluation. In
ically addresses the various phases of a mining cycle. This map firstly addition to AHP’s extensive use within and outside the mining industry,
addresses the six mining phases, thereafter each phase was further its ability to model complex decision problems makes it ideal (Musing­
expanded to capture value drivers that impact a mine operation. The six wini and Minnitt, 2008; Sitorus et al., 2019). PROMETHEE is best known
mining phases are exploration; project evaluation and planning; mine for its unique ability to sort through decision alternatives and rank the
design and construction; operations; decommissioning/closure; and post result in order of preference without eliminating any options (Sitorus
closure. The seven main value driving pillars under which other value et al., 2019; Yuen and Ting, 2012).
drivers were categorized include production; supply chain; profitability
and cost control; productivity and asset efficiency; mineral resource 2.1. Case study and scenario description
management; socio-economic factors; and health, environment, safety
and legal. Overall, about 550 technologies were covered in the map. In this study, a hypothetical case study was used to model an ideal
The proposed decision algorithm’s ability to solve decision problems situation in a mining company. A single decision maker decided on
with both fuzzy and precision-based (non-fuzzy) characteristics makes it behalf of a company to select the best technology to adopt to improve its
ideally suitable for real life mine decision-making scenarios. This is overall mining operation, while also keeping in alignment the strategic
because, most times, while selecting technologies or evaluating them for goal of the company.
a proposed project, the information available to make those decisions Mining Company A mines iron ore through open pit mining method
are usually imprecise and uncertain. The general-purpose nature of the and has in the past been faced with safety challenges. In its previous
analytical procedures also makes it applicable across all mining fiscal years, the management has decided to opt for adopting technol­
methods, processes, and operations. However, in establishing the ogies to help improve this. The main goal for management is to ensure
context of this study, the focus was on surface mining operations. The that the open pit mine’s hazard and risk is better managed and antici­
Multi-Attribute Decision Making (MADM) methodology of operation pated to ensure maximum safety. In line with this, the management
research was used as the basis for the analytical procedure. team has prioritized adopting Hazard Identification technologies from
The fusion of analytic hierarchy process (AHP) and preference Jacobs (2016) technology map. From the six technology available op­
ranking organisation method of enrichment (PROMETHEE) techniques tions, the goal is to select the most optimal technology for the surface
were employed in this study. The AHP was used to determine the hier­ mine operation and by adopting it, reduce the occurrences of
archal weight and its consistency, while the PROMETHEE method was life-threatening happenings by 50% in 3 years. The budget for this
used to carry out the overall process evaluation. Within the mining in­ technical upgrade is about 60,000 currency units (C.U).
dustry, AHP and PROMETHEE MCDM techniques have been used across
the mining value chain to solve various operational constrains. Some of 2.2. The conceptual framework
them include plant and equipment selection (Wang and Tu, 2015;
Owusu-Mensah and Musingwini, 2011); mine site selection (Dey and For this study, the procedures depicted in Fig. 1 shows a summary of
Ramcharan, 2008; Hudej et al., 2013); mine method selelection (Owu­ the process undertaken to achieve the goal of sorting and selecting the
su-Mensah and Musingwini, 2011; Samimi et al., 2009); and mineral most preferred technology suitable to meet the mining company’s
processing (Savic et al., 2015). Some of the other industries where organisational goal.
MCDM has been used are real estate; information, communication and Fig. 1 breaks down the general steps needed in the MCDA process
technology (ICT); finance; manufacturing; economics; management; and computation. Stemming out from this, the appropriate framework used
environment (Sitorus et al., 2019). In addition, the outcome of the study to solve the decision problem in this study is drawn and shown as out­
is critical in forming the necessary technology business cases needed for lined in Fig. 4. In Fig. 2, box A is the expansion of the 4th step in Fig. 1
the decision makers. and it shows the process and method by which the weights of the
decision-making criteria were obtained. Here, the AHP MCDM method is
2. Methods used. Similarly, Box B is an expansion of the 5th step in Fig. 1 and it
shows the process and methods used in the computational evaluation of
The complexity of operations within the mining industry makes it the judgement matrix. The PROMETHEE II was used to make the
important to take a multifaceted approach to solving the managerial, calculation. In addition, fuzzy set theory was used to calculate the
technical, and technological problems. Solving these problems, howev­ evaluative values for the objective criteria.
er, involves the use of multidisciplinary knowledge which criss-crosses
economics, environments, politics, policies, and social aspects (Sitorus 2.3. MCDM approach
et al., 2019). One of the ways the mining industry has engaged with such
is by generating several workable solutions for such problems, there­ According to Vujic et al. (2013), the quality of a decision depends on
after, selecting the most suitable one. the quality of the criteria and the alternatives, and to a certain extent the

2
O. Dayo-Olupona et al. Resources Policy 69 (2020) 101879

Fig. 1. General MCDA process framework.

selection method. Furthermore, selecting of the appropriate 1980’s (Koksalan et al., 2011). It is a method that is particularly skilled
mathematical-model approach hinges on the problem type and struc­ in outranking results of the multicriteria problem. Using the decision
ture, and the decision maker’s proficiency. In the following sections, the makers criteria and preferences, the outranking method does not elim­
structure, alternatives, and criteria for the multicriteria decision prob­ inate any alternatives in the pairwise comparison; rather, it arranges the
lem are addressed. The techniques of MCDM used in this study to solve results in order of preference.
this problem was a fusion of AHP and PROMETHEE. Generally, the AHP, Abdullah et al. (2019) summarized the computational procedure of
PROMETHEE and Elimination Et Choix Traduisant la Realité (ELECTRE) PROMETHEE into seven steps. These steps were adopted for the
are some of the most frequently used MCDM techniques (Sitorus et al., computation of the data in this study; however, they were modified and
2019). As reported by Sitorus et al. (2019), these three techniques have compressed to a total of five steps.
unique characteristics needed to tackle different types of challenges. The
AHP is used to solve the choice challenge, the PROMETHEE to solve the Step 1: Determine the set of possible alternatives in the decision
ranking challenge and ELECTRE can solve both the ranking and sorting problem.
challenges. In the following sections the stepwise procedures of the two Step 2: Determine the criteria
techniques used in this study are described. Step 3: Determine the weights for each of the criteria
Step 4: Construct the decision matrix
2.3.1. Analytic hierarchy process (AHP) Step 5: Determination of the preference order. It involves;
AHP is an MCDA method proposed by Thomas L. Saaty, an operation • Normalization of the decision matrix;
research scientist in 1977 (Koksalan et al., 2011). It focuses on struc­ • Determination of the deviation between the decision criteria;
turing complex multi criteria decision problems into a • Definition of the preference function; and
multi-dimensional hierarchical model. The AHP technique usually en­ • Determination of the multicriteria/aggregated preference index.
tails the following three steps:
The formula and processes used in carrying out all the outlined steps
a) The Hierarchical Judgment Matrix Design; in step 5 are expanded in Abdullah et al. (2019).
b) The Hierarchical Relative Weight calculation; and
c) The Judgement Consistency Step. 3. Computational evaluation

The decision to include AHP as one of the two techniques used in this In this section, the detailed calculations supporting the evaluation
study were based on three reasons: steps listed in section 2.3.2 are discussed. In addition, critical back­
ground and scoping terms are explained.
1. The AHP method detects inconsistency judgements and supplies an
estimate degree of inconsistency (Musingwini and Minnitt, 2008);
3.1. Alternatives
2. The AHP is one of the most prominently used MCDA methods, and its
application in the mining industry is becoming pre-eminent (Mus­
Step 1: Determine the set of possible alternatives in the decision
ingwini and Minnitt, 2008); and
problem.
3. The AHP is categorized as one of the techniques that are best used to
solve decision problems (Haddad and Sanders, 2018; Stojanović
According to Triantaphyllou (2000), alternatives are the choices of
et al., 2015).
action made available to a decision maker. For this study, the set of
alternatives were the technology options and they are finite in number.
2.3.2. Preference ranking organisation method for enrichment evaluation
They were screened, prioritized, and eventually ranked.
(PROMETHEE)
From Jacobs (2016) technology map, as depicted in Fig. 3, the
The PROMETHEE method was proposed by Jean-Pierre Brans in the
management team chose a phase within the value chain for which the

3
O. Dayo-Olupona et al. Resources Policy 69 (2020) 101879

Fig. 2. Schematic framework for technology adoption.

4
 O. Dayo-Olupona et al.
5

Resources Policy 69 (2020) 101879

Fig. 3. Mining Cycle Framework Highlighting the Procedure for selecting Technology Alternatives from the Technology Map.
O. Dayo-Olupona et al. Resources Policy 69 (2020) 101879

technology upgrade is targeted; the operations (mine-to-mill) phase was These weights are normalized to sum up to one as shown in Equation (1)
chosen. Thereafter, the main value driving pillar was identified from the (Abdullah et al., 2019).
left vertical axis: Health, Environment Safety and Legal were selected in
∑
n
this case. The management team then identified the Hazard Identifica­ Wj = 1 (1)
tion sub-category under which there are six technology options. They j=1

are:
The weights were arrived at by following procedure 1 to 3 of the AHP
[AR, AA, AI, UAVs, IoT and IoE].
technique described in Section 2.3.1. On comparing the criteria with
each other in a pairwise comparison matrix, a hierarchical judgement
• Augmented Reality (AR) A1;
matrix for the criteria was formed as shown in Table 3. The scale used
• Advanced analytics (AA) A2;
here to make this subjective pairwise comparison was recommended by
• Artificial Intelligence (AI) A3;
Saaty (1995). Saaty (1995) referred to it as “the fundamental scale of the
• Unmanned Air Vehicles (UAVs) A4;
AHP with absolute values of 1–9”. Table 2 shows the table of scale. Each
• Internet of Things (IoT) A5; and
number corresponds to the preference strength of one variable or
• Internet of Everything (IoE) A6.
element over another. “1” shows that the elements to be compared are of
equal importance, while “9” show that one of the elements is extremely
3.2. Criteria more important than the other. In situations where a compromise is
needed between both the elements to be compared, the intermediate
Step 2: Determine the criteria. values of 2, 4, 6 and 8 would be used.
Criteria are sometimes known as “goals” or “decision criteria” or
“attributes”. They can be described as a measure of effectiveness and 3.3.1. The Hierarchical Relative Weight calculation
they form the basis on which the evaluation is made. (Hwang and Yoon, In order to meet the precision requirement for calculating the rela­
1981). This study extracted its decision-making criteria from five pub­ tive weight using the AHP technique, the approximate calculation was
lished articles across four industries. The industries involved are mining, done using the square root method. To calculate the geometric means of
technology, manufacturing, and aerospace (Ordoobadi, 2012; Taha the varying elements on each row of the Judgement Matrix, Equation (2)
et al., 2011; Poulin et al., 2013; Vujic et al., 2013; Stojanović et al., (Wang and Tu, 2015) was used.
2015). Bartos (2007) compared the mining industry’s rate of technology √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
̅
∏n
adoption to manufacturing, and the semiconductor (technology) in­ ϖi = n aij , i = 1, 2, …, n, (2)
dustry. Drake et al. (2017) notably stated that there is no rule governing
j=1

the number of criteria to be included in an analysis. Furthermore, the ϖi

higher the number of criteria, the higher the complexity and cognitive ω = ∑n , i = 1, 2, …, n, (3)
i=1 ϖ i
effort needed for the decision maker to fill out the judgment matrix.
Based on this, the decision maker chose only four criteria (as displayed where aij stands for the elements of the judgment matrix, and ϖ = (ϖ i ,
in Table 1) with a minimum of one under a major category. The four ϖ 2 , …, ϖ n ). Thereafter ϖ i is normalized to give the weight vector ω, the
chosen criteria are: normalization was done using Equation (3) (Wang and Tu, 2015). The
resulting weight vector ω = (ω1 , ω2 , …, ωn ) is the relative weights of
a) Economic analysis; each criteria with respect to the goal. This is represented as follows:
b) Suitability;
ω = (0.432, 0.075, 0.114, 0.380)
c) Strategic fit; and
d) Operational safety. The “Economics” criterion held the largest weight with the value of
0.432, this was closely followed by the “Operational Safety” criterion
Fig. 4 also shows the hierarchal structure of the technology decision weighing at 0.380. The “Strategic fit” and “Efficiency” criteria respec­
problem that was computed in this study. This gives a pictorial view of tively have the values of 0.114 and 0.075, respectively.
all the highlighted criteria and alternatives highlighted in Sections 3.1
and 3.2 and its relation to the overall goal of the decision maker. The 3.3.2. The Judgement Consistency Step
framework shows the goal of the decision problem which is to select the The checking of consistency in AHP judgment matrices is usually
most optimal technology for the Iron ore mine, the four subjective and done to measure its credibility. According to Saaty (1995), a pairwise
objective criteria and the six technology alternatives (A1 – A6) from comparison in an evaluation matrix is said to be consistent when the
which the decision was made. consistency ratio is smaller than 10%. This was calculated by first
As shown in Fig. 4, the economic analysis criterion was the objective computing the Consistency Index (Table 4). Next, the consistency index
criterion, while the other three were the subjective criteria. value was divided by the correction value know as random consistency
index (RCI) to get the Consistency Ratio (Table 4).
3.3. wt As shown in Table 4, the Consistency Ratio of the AHP matrix is
0.09645, which is less than the 0.1 (10%) standard. This translates to the
Step 3: Determine the weights for each of the criteria fact that the AHP judgement matrix is consistent;
∑n
(Rϖ)i
Weights are assigned to each criterion to measure its individual λmax = (4)
nϖ i
importance in relation to the general objectives of the decision problem. i=1

The weight of a criterion is usually denoted by the symbol Wj where

where RCI is the correction value and is known as the Random Consis­
Wj > 0 and always within a set of real numbers (Triantaphyllou, 2000).
tency Index (Table 5), depending on the dimensions of the matrices; λmax
represents the maximum eigenvalue (Equation (4)) (Wang and Tu,
Table 1
2015).
Criteria selection.
where R is the judgment matrix and (Rϖ)i is the ith element of the
Economic Factors Technical Factors Ergonomic Factors matrix. (Rϖ)i
Economic analysis Suitability Strategic fit The RCI value for the order of matrix table shown in Table 5 is a table
Operational safety of standard randomized values that are used while making AHP

6
O. Dayo-Olupona et al. Resources Policy 69 (2020) 101879

Fig. 4. Mcda framework for technology selection (adapted from Wang and Tu, 2015).

calculations.
Table 2
Measurement scale of relative importance (Saaty, 1995).
Verbal judgement or preference Numerical
3.4. Judgment matrix
rating
Step 4: Construct the decision matrix.
Equally Important 1
Moderately Important 3
Strongly Important 5 For all MCDM methods, a decision matrix, otherwise known as
Very Strongly Important 7 judgement matrix or decision table forms the base structure from which
Extremely Important 9
a pairwise comparison is made. It is illustrated as follows:
Intermediate values between two adjacent Judgment (when 2, 4, 6, and 8
compromise is needed) Suppose a problem has m alternatives with a set of alternatives a =
[A1, A2, …, Am ] and n number of decision criteria with the set of criteria.
X = [x1 , x2 , … , xn ]xij (i = 1, 2, …, m; j = 1, 2, …, n) are the elements of
the matrix which denotes the performance value of the ith alternative Ai
Table 3
Pairwise comparing matrix for criteria. in terms of the jth criterion ((Equation (5)) (Wang and Tu, 2015).
⎡ ⎤
Economic Efficiency Strategic Fit Operational safety x11 x12 ⋯ x1n
( ) ⎢ x21 x22 ⋯ x2n ⎥
Economic 1 5 5 1 R = Xij m×n = ⎢ ⎣ ⋮
⎥ (5)
Efficiency 1/5 1 1/3 1/3 ⋮ ⋱ ⋮ ⎦
Strategic Fit 1/5 3 1 1/5 xm1 xm2 ⋯ xmn
Operational safety 1 3 5 1
⎡ ⎤
Economic Efficieny Strategic Fit Operational Safety
⎢ AR x11 x12 x13 x14 ⎥
⎢ ⎥
⎢ AA x21 x22 x23 x24 ⎥
Table 4 ( ) ⎢ ⎥
R= Xij m×n = ⎢
⎢ AI x31 x32 x33 x31 ⎥
Judgement consistency computational table. ⎢ UAVs
⎥
⎥
⎢ x41 x42 x43 x53 ⎥
Equation Calculated Value ⎣ IoT x51 x52 x53 x54 ⎦
Consistency Index (CI) λmax − n 0.08681 IoE x61 x62 x63 x64
n− 1 (6)
Consistency Ratio (CR) 1 0.09645
= × CI
RCI After including the alternatives and criterions, the judgment matrix
1 λmax − n
=
RCI
×
n− 1 looks as follows (Equation (6));
As shown in the matrix 6, alternatives a = [A1, A2, …, Am ] are
[AR, AA, AI, UAVs, IoT, IoE]
while criteria X = [x1 , x2 , … , xn ] are. [Economic, Efficency,

7
O. Dayo-Olupona et al. Resources Policy 69 (2020) 101879

Table 5
RCI values for different Order of Matrix (n) (Wang and Tu, 2015).
Order of Matrix (n) 1 2 3 4 5 6 7 8 9

RCI value 0.00 0.00 0.58 0.90 1.12 1.24 1.32 1.41 1.45

Strategic Fit, Operational Safety].

Table 7
For this study, the subjective criteria; suitability, strategic fit and
Decision matrix with filled subjective criteria column.
operational safety criteria are qualitative and were described by lin­
guistic assessments. Contrarily, the objective criterion - the economic Economic Efficiency Strategic Fit Operational safety
analysis - is quantitative and was appraised in monetary terms. AR x11 1 5 3
For the subjective criteria, filling in the decision matrix was based on AA x21 3 3 3
the linguistic numeric scale as presented in Table 6. This was based on AI x31 3 3 2
the subjective knowledge of a decision maker. Filling up the matrix UAVs x41 4 4 4
required the decision maker to answer questions such as, “how much IoT x51 3 4 3
value will UAVs add to operational safety of the mining project”? As IoE x61 3 4 3
shown in Table 7, the decision maker answered by stating that the po­
tential value add for UAVs to operational safety is high. This is assigned
the number 4 in the fourth row, fourth column or (4,4). Notably, “effi­
ciency” and “strategic fit” criteria approached the maximum, while the
“operation safety” criterion approached the minimum.
For the objective criteria, fuzzy set theory would be used to get the
crisp quantitative net present value (NPV) values to be used in the
judgment matrix and it approached the maximum.

3.4.1. Fuzzy set theory

Chan et al. (2000) highlighted that in order to evaluate the economic
factor for technology selection several researchers have adopted
precision-based methods. Some of them include return on investment
(ROI), net present value (NPV), payback period (PB) and the internal
rate of return (IRR). However, the crisp nature of the data required for
factors such as investment cost, interest rate, salvage value, gross in­
come, depreciation, which are quite difficult to quantify.
In such circumstances, decision makers usually give assessment
based on practical knowledge, and subjective judgement.
In a bid to convey estimations, linguistic terms such as “about
$10,000”, “high”, “approximately between $100,000 and $250,000”, Fig. 5. Triangular fuzzy number (Wang and Tu, 2015).
“around 25%“, “very low” are used by decision makers. Chan et al.
(2000), however, introduces the fuzzy set theory to deal with the ( ) ( )
vagueness of human thought. Tackling these ambiguities associated Xmj = Gmj − Cmj − Gmj − Cmj − Dmj × Tm − Km + Lmj + Vmj (7)
with the linguistic estimations involves converting these linguistic terms
where.
to fuzzy numbers. From triangular fuzzy number model illustrated in
Fig. 5, f(x) is the membership function of fuzzy number x, where x =
Xmj is the net total cash flow of the technology m at the end of year j;
(m, a, b) i.e. m − a, < x < m + b and f(x) ∈ [0, 1].
Within the scope (m − a, m), f(x) is a linear increasing monotone Gmj is the turnover of the technology m at the end of year j;
function, while for (m, m +b), f(x) is a linear monotone decreasing Cmj is the operating expenses of the technology m; at the end of the
function (Chan et al., 2000), where, m − a will serve as the lower year j;
boundary and m + b will be the upper boundary. Therefore, (m, a, b) = Dmj is the depreciation amount of technology m in year j;
(m − a, m, m + b). Tm is the tax rate of the technology m;
In this study, the economics analysis criteria of the multicriteria Km is the investment cost of the technology m;
problem could not be quantified accurately, therefore, the fuzzy set Lmj is the salvage value received in year j;
theory was introduced to help quantify the criterion. It was calculated by Vmj is the incremental tax credit of the technology m in year j.
the fuzzy cash flow model proposed by Komolavanij (1995). The oper­
ations of a fuzzy number can also further be obtained from (Komola­ Thereafter, the NPV of m technologies will be calculated using
vanij, 1995). (Equation (8));
The proposed equation was ((Equation (7)); j
∑ Xmj
NPVm = (8)
n=0
(1 + i)n
Table 6
Linguistic scale for subjective criteria. where.
Linguistic Scale Numeric Scale
NPVm is the net present value of technology m;
Very High VH 5
i is the discount rate; and.
High H 4
Medium M 3 j is the life of project.
Low L 2
Very Low VL 1

8
O. Dayo-Olupona et al. Resources Policy 69 (2020) 101879

3.4.2. Fuzzy economic data

The following hypothetical economic data and assumptions struc­ Step 5: Determination of the preference order
tured after the Komolavanij (1995) format were used for the economic
fuzzy NPV calculations: Here the choice is usually made between the PROMETHEE I and the
PROMETHEE II. This is done to determine if the result would be ranked
• The budget for this technical upgrade is about 60,000 currency units partially or completely. PROMETHEE I rank the resulting alternatives
(C.U); partially, while PROMETHEE II ranks them completely. For this study,
• The gross income/turn over (Gmj ) are assumed to be about 50% of the the complete ranking of PROMETHEE II was employed and Equation
investment cost on a yearly basis; (11) was used;
• Operating expenses (Cmj ) are assumed to be about 20% of the in­
ϕ(a) = ϕ+ (a) − ϕ− (a) (11)
vestment cost;
• Tax rate (Tm ) of technology is assumed to be 40% based on Komo­
where, ϕ(a) represents the net outranking flow for each alter­
lavanij (1995);
native, ϕ+ (a) means positive outranking flow or leaving flow. This also
• The depreciation (Dmj ) amount of technology is assumed to be about
means how a dominates other alternatives. ϕ− (a) means the negative
33.3% based on (Ernst and Young, 2018) Discount rates (i) are var­
outranking flow or entering flow. This also shows how a is been domi­
iable and presented in Table 9;
nated by other alternatives.
• Lmj = 0 (Komolavanij, 1995); and
The following also shows reference relations of the alternatives.
• Vmj = 0 (Komolavanij, 1995).
( )
a outranks of b aP(II) b i ff ϕ(a) > ϕ(b), ∀a, b ∈ A
Lastly, the decision maker’s assumed fuzziness is said to be within
( )
the 7% range due to the uncertainty in the data. The data provided a outranks of b aP(II) b i ff ϕ(a) = ϕ(b), ∀a, b ∈ A
above is presented and expanded in Table 9. It provides the basis to carry
out the fuzzy calculations to evaluate the annual cash flow for each of Thus, based on the ϕ(a) values of each alternatives, all alternatives
the technology options. All the variables present in equations (7) and (8) can be compared. The alternatives with the highest ϕ(a) values repre­
were specified as a triangular fuzzy number. They represent, “the sent the most preferred.
possible value”, “the pessimistic value” and “the optimistic value”. In Table 10 shows the computed leaving flow values, entering flow
evaluating the objective criterion, the decision-maker analysed the fuzzy values and net outranking flow values for each of the technology al­
cash flow analogous to the technology alternatives A1, A2, A3, A4, A5, ternatives. Additionally, the last column of the row shows the overall
and A6 for the purpose of calculating the fuzzy NPV. rank for the technology option.
From Table 8, it is noticeable that year 0 is the investment year and
the subsequent years are the years where returns were derived from the 4. Results
investment. For each of the technology options, the Net cash flow value
(Xmj ) were calculated. This was done for year 0 and the subsequent The results of the case study considered are represented in Table 10
three-year period. This cash flow values were presented in the fuzzy and Fig. 6. Table 10 shows the net flow values of each of the technology
pessimistic, possible and optimistic value. Subsequently, the fuzzy NPV alternatives and the respective ranks when arranged from the highest
values were calculated. These data are presented in Table 9. The in-line (best option) to the lowest (worse option). Fig. 6 shows the pictorial
computational procedure that generated the fuzzy economic values is hierarchical order of the results shown in Table 10 as generated from the
displayed in the Appendix. PROMETHEE-GAIA (graphical analysis for interactive assistance) soft­
However, the fuzzy economic values cannot be used directly in the ware. This, in summary, indicates that AI (A3) is the best technology fit
decision matrix. Goumas and Lygerou (2000) proposed the use of the for the iron ore mining project in comparison with the other five tech­
Yanger index function (Equation (9)) to “defuzz” the fuzzy numbers in nology options. In order of preference, the technology can be repre­
order to get crisp (non-fuzzy) numbers that can be used in the decision sented as follows;
matrix. A3, > A4 > (A5 = A6) > A2 > A1
In this case, the fuzzy NPV values (shown in Table 9) were “defuz­
zied” and inputted in the matrix shown in Equation (10). This was done where ‘>’ means ‘is more preferred’ and ‘ = ’ means is ‘equally preferred
by using Equation (9) (Wang and Tu, 2015); to’.

(3m − a + b)
f (x) = F(m, a, b) = (9) 5. Discussion and conclusion
3
The resultant crisp NPV values is shown in Table 10. Thereafter, The contextualizing of the authors result with previous studies may
these values were inserted in the decision matrix. be quite a challenge, since it was discovered that limited literature
addressed the application and adoption of technology in mining. Addi­
tionally, further research also revealed that only two studies within the
3.5. The decision matrix
mining industry employed the use of the MCDM techniques and methods
to adopt other forms of technology. The first was Stojanović et al. (2015)
The resulting decision matrix is presented as follows:
and the second was Vujic et al. (2013).
Economical Considering the methodology used, Vujic et al. (2013) used only the
Efficiency PROMETHEE II MCDM technique in comparing the mode of machinery
Strategic fit arrangement. Similarly, Stojanović et al. (2015) used the combined AHP
Operational⎡ safety ⎤ and ELECTRE method in selecting the machinery operation method. The
AR 2569.58 1 5 3 procedural calculation for the AHP and PROMETHEE II methods used in
(10)
AA ⎢ ⎢ 3481.89 3 3 3⎥⎥ Stojanović et al. (2015) and Vujic et al. (2013) respectively, are similar
( ) AI ⎢⎢ 6415.20 3 3 2⎥
R = Xij m×n = ⎥ to the one used in this study, however, the nature of technology
UAVs ⎢⎢ 8212.89 4 4 4⎥
IoT ⎣ 4887.77 3 4 3⎦
⎥ considered and the nature of data used are completely different.
IoE 4887.77 3 4 3 The current study included sensitivity analysis (section 5.2). In
addition to the description provided in the section, another reason why

9
 O. Dayo-Olupona et al.
Table 8
Economic data of technology A1, A2, A3, A4, A5, A6.
Technology A1 Technology A2 Technology A3 Technology A4 Technology A5 Technology A6

AR AA AI UAV IoT IoE

Pessimistic Possible Optimistic Pessimistic Possible Optimistic Pessimistic Possible Optimistic Pessimistic Possible Optimistic Pessimistic Possible Optimistic Pessimistic Possible Optimistic
Value Value Value Value Value Value Value Value Value Value Value Value Value Value Value Value Value Value

At the end of year 0

G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
C 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
D 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
T 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40%
K 9300.0 10000.0 10700.0 12090.0 13000.0 13910.0 19530.0 21000.0 22470.0 22320.0 24000.0 25680.0 14880.0 16000.0 17120.0 14880.0 16000.0 17120.0
L 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
V 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
i 12% 12% 12% 11.5% 11.5% 11.5% 8% 10% 12% 6% 8.5% 11% 8% 10% 12% 8% 10% 12%

At the end of year 1

G 7905.0 8500.0 9095.0 10276.5 11050.0 11823.5 16600.5 17850.0 19099.5 18972.0 20400.0 21828.0 12648.0 13600.0 14552.0 12648.0 13600.0 14552.0
C 1860.0 2000.0 2140.0 2418.0 2600.0 2782.0 3906.0 4200.0 4494.0 4464.0 4800.0 5136.0 2976.0 3200.0 3424.0 2976.0 3200.0 3424.0
D 3100.0 3333.3 3566.7 4030.0 4333.3 4636.7 6510.0 7000.0 7490.0 7440.0 8000.0 8560.0 4960.0 5333.3 5706.7 4960.0 5333.3 5706.7
T 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40%
K 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
10

L 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
V 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
i 12% 12% 12% 11.5% 11.5% 11.5% 8% 10% 12% 6% 8.5% 11% 8% 10% 12% 8% 10% 12%

At the end of year 2

G 7905.00 8500.0 9095.0 10276.5 11050.0 11823.5 16600.5 17850.0 19099.5 18972.0 20400.0 21828.0 12648.0 13600.0 14552.0 12648.0 13600.0 14552.0
C 1860.0 2000.0 2140.0 2418.0 2600.0 2782.0 3906.0 4200.0 4494.0 4464.0 4800.0 5136.0 2976.0 3200.0 3424.0 2976.0 3200.0 3424.0
D 3100.0 3333.3 3566.7 4030.0 4333.3 4636.7 6510.0 7000.0 7490.0 7440.0 8000.0 8560.0 4960.0 5333.3 5706.7 4960.0 5333.3 5706.7
T 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40%
K 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
L 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
V 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
i 12% 12% 12% 11.5% 11.5% 11.5% 8% 10% 12% 6% 8.5% 11% 8% 10% 12% 8% 10% 12%

At the end of year 3

G 7905.0 8500.0 9095.0 10276.5 11050.0 11823.5 16600.5 17850.0 19099.5 18972.0 20400.0 21828.0 12648.0 13600.0 14552.0 12648.0 13600.0 14552.0
C 1860.0 2000.0 2140.0 2418.0 2600.0 2782.0 3906.0 4200.0 4494.0 4464.0 4800.0 5136.0 2976.0 3200.0 3424.0 2976.0 3200.0 3424.0
D 3100.0 3333.3 3566.7 4030.0 4333.3 4636.7 6510.0 7000.0 7490.0 7440.0 8000.0 8560.0 4960.0 5333.3 5706.7 4960.0 5333.3 5706.7
T 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40% 40%

Resources Policy 69 (2020) 101879

K 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
L 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
V 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
i 12% 12% 12% 11.5% 11.5% 11.5% 8% 10% 12% 6% 8.5% 11% 8% 10% 12% 8% 10% 12%
O. Dayo-Olupona et al. Resources Policy 69 (2020) 101879

sensitivity analysis is usually conducted is to review and understand the

behaviour of the computed results (Haddad and Sanders, 2018). For this

Optimistic

− 14880.0

8902.14
reason, the last paragraph of section 5.2 discussed the nature and

9228.3
9228.3
9228.3
Value
behaviour of this study result in comparison with other studies.

− 16000.0
5.1. GAIA plane analysis

4823.24
Possible

8373.3
8373.3
8373.3
Value
Technology A6
Fig. 7 shows the relationship between criteria and the technology

Pessimistic
alternatives and the criteria. Using a 90% representative value, this

− 17120.0

4887.77
GAIA plane analysis will help to understand this relationship. It is

7518.4
7518.4
7518.4
937.93
Value
important to point out that the light-blue square boxes are the tech­

IoE

IoE
nology alternatives, while the vectors with blue lines and rhombuses are

Optimistic

− 14880.0
the criteria.

8902.14
9228.3
9228.3
9228.3
Value
The orientation of these criterial vectors shows how closely related
or conflicting one criterion is to another. From Fig. 7, the “Economic”

− 16000.0
and “Efficiency” criteria are closely related. This means technologies

4823.24
Possible

8373.3
8373.3
8373.3
with economic values also have high efficiency rates, while “Operational

Value
Safety” and “Strategic Fit” are relatively conflicting with themselves and

Technology A5
other criteria. The decision vector is the thick red axis with a big dot. Its

Pessimistic

− 17120.0

4887.77
orientation indicates the degree of alignment of each criterion with the

7518.4
7518.4
7518.4
937.93
Value
PROMETHEE ranking. The decision axis is opposite to strategic fit;

IoT

IoT
therefore, it is expected to find technologies with higher economic NPVs

Optimistic

− 22320.0
and higher efficiency at the top of the PROMETHEE rankings.

14680.90
13842.4
13842.4
13842.4
The technologies that are located close to a criterion show they are

Value
particularly good at that criteria. For instance, the technology AR, IoEs
and IoTs (A1, A5, A6) are situated close to the “Strategic Fit” criteria.

− 24000.0
12560.0
12560.0
12560.0
8078.52
Possible
This means that these technologies have the best strategic fit to the

Value
mining project. For the “Operational Safety” criteria, Technology AA
Technology A4

(A2) has the best fit. Similarly, UAVs (A4) is independently the most
Pessimistic

− 25680.0
efficient and has economic value.

11277.6
11277.6
11277.6
1879.24

8212.89
Value
UAVs

UAVs
In the same way, technology alternatives with similar profiles are
situated close to each other as in the case of IOEs and IOTs (A5, A6).
Their profiles are similar to the extent where they literally overlap. On
Optimistic

− 19530.0

11684.06
12112.1
12112.1
12112.1
the other hand, in the same Fig. 7, UAVs and AA (A5, A2) tend to have
Value

differing profiles because, they are situated very far from each other.
− 21000.0
10990.0
10990.0
10990.0
6330.50
Possible

5.2. Sensitivity analysis

Value
Technology A3
Cash Flow Model Showing the Fuzzy and Crisp NPV values for the Technology Alternatives.

Sensitivity analysis in this study aims to find out how much the
Pessimistic

output of the model used to solve the decision problem is affected by the
− 22470.0

1231.03

6415.20
9867.9
9867.9
9867.9

uncertainty in the input variable. This is carried out to identify any close
Value

competition to the most preferred alternative. The validation process

AI

AI

involves analysing the effect of weight modification for each criterion on

Optimistic

− 12090.0

6074.72

the technology choice for the mine which is shown in Fig. 8.

7498.0
7498.0
7498.0
Value

For each of the Graphs (A-D) in Fig. 8, the vertical axis is the net flow
axis, while the horizontal axis is the criterion weight axis labelled from
− 13000.0

0% to 100%. Each of the light-blue lines shows how each of the tech­
3481.89
Possible

6803.3
6803.3
6803.3
Value

nology alternatives change with a modification of the criterion values.

Technology A2

It is important to note that the green and red vertical lines make it
easy to identify the technology alternative ranking at the weight
Pessimistic

− 13910.0

3481.89

assigned in this study for the criteria considered in the graph. For
6108.7
6108.7
6108.7
889.06
Value

instance, the AI (A3) is ranked at the top for the 11% weight for the
AA

AA

strategic fit criteria (Graph A), 43% weight for economic criteria (Graph
Optimistic

B), 38% weight for operational safety criteria (Graph C), and 7% weight
− 9300.0

4552.96
5767.7
5767.7
5767.7
Value

for efficiency criteria (Graph D).

The Graph (A-D) of Fig. 8 answers the question “what is the effect of
− 10000.0

assigning more or less weight to each of the criteria”. In Graph D, the

2569.58
Possible

5233.3
5233.3
5233.3

efficiency criterion is considered, and for the criterion weight lesser than
Value

or equal to 34.19% (0.34), the AI (A3) technology option was ranked at

Technology A1

the top of the alternatives. However, for the efficiency criterion weight
Pessimistic

− 10700.0

greater than 34.19%, UAVs (A4) were ranked the first.

Crisp Net 2569.58
4699.0
4699.0
4699.0
586.21
Value

In Graph C, the operational safety criterion is considered and for

AR

AR

criterion weights lesser than or equal to 22.39% (0.22) UAVs were

Fuzzy Net
Present

Present

ranked first choice among all the technology alternatives. However, for
Value

Value
Table 9

Year 0
Year 1
Year 2
Year 3

operational criterion weight greater than 22.39%, AI (A3) was ranked

first.

11
O. Dayo-Olupona et al. Resources Policy 69 (2020) 101879

Table 10
The flow values.
ϕ+ (a) ϕ− (a) ϕ(a) Rank

AR 0.0980 0.2690 − 0.1710 6

AA 0.0516 0.2168 − 0.1652 5
AI 0.3237 0.0744 0.2493 1
UAVs 0.2880 0.1995 0.0885 2
IoT 0.1064 0.1072 − 0.0008 3
IoE 0.1064 0.1072 − 0.0008 3

Fig. 7. The GAIA plane analysis of the decision problem.

and the result. The original calculated criteria-weights assigned by the

decision maker for technology decision problem considered in this study
is depicted by the deep-blue bars in Fig. 9. The six bars highlighted in
light-blue colour in Fig. 9 show the technology alternatives, arranging
them from the most preferred to the least suitable.

5.2.1. Behaviour of result

To test for sensitivity, all the weights are set equal and the results
displayed by the light-blue bar charts in the upper portion of Fig. 10
which shows that UAVs (A4) was the most preferred and AA (A2) was the
least preferred. These noted variances show that the applied model used
in this study is moderately stable as it is sensitive. An extremely stable
model would not have generated a different alternative rank or result
even when the criteria are varied or set equal. Similar studies with
outcomes that were sensitive and unstable in nature after the sensitivity
analysis which were carried out incudes the studies by Dey and Ram­
charan (2008) and Haddad and Sanders (2018). On the other hand,
studies with stable outcomes include: Wang and Tu (2015) and Talukder
and Hipel (2018). The study by Thomaidis et al. (2006) suggests that
MCDM models with crisp weights and rates have final ranking that are
quite sensitive. Conversely, MCDM models with fuzzy set weights and
rates are less sensitive to exact assessments and this is due to the fuzzy
set which allows for the interpolation of a range of values. In some way,
fuzzy set in itself, performs a second-order sensitivity analysis (Tho­
maidis et al., 2006). In this study, a combination of crisp and fuzzy
variables was used, however, the crisp features were more than fuzzy.
Fig. 6. The complete ranking of the technology alternatives flow.
This makes a case for, why the nature of the result is said to be
moderately sensitive. Other possible reasons why this problem model is
In Graph B, the economic criterion is considered and for weights sensitive are listed below:
lesser than or equal to 72.14% (0.72), AI (A3) was ranked at the top,
However, if weight value higher than 72.14% was assigned to the eco­ 1. The presence of more subjective criteria than objective criteria;
nomic criterion, UAVs (A4) became the first on rank. 2. A single decision maker was used in this study, making it not the
In Graph A, the strategic fit criterion is considered and for weight perfect representation for a decision-making situation in a mine; and
lesser than or equal to 37.05% (0.37), AI (A3) was ranked in first place. 3. The hypothetical nature of the case study.
For the efficiency criterion weight between 37.05% and 47.05%, UAVs
(A4) were ranked in first place. If the weight value was greater than
47.05%, and was assigned to the same criterion, AR (A1) was ranked in 5.3. The decision
the first place.
Similarly, using the walking weights interactive feature of the Based on the need of the company, the most preferred technology
PROMETHEE-GAIA software, the sensitivity of the result to variability in would be AI (A3). Considering the budget of 60,000 C.U set aside by the
the criterion weighing is tested. Fig. 9 and Fig. 10 show the weightings mining company, an additional technology can be adopted in addition to

12
O. Dayo-Olupona et al. Resources Policy 69 (2020) 101879

Fig. 8. Criterion weight variability graphs.

A3. Based on this, technology (A4) – Unmanned Air Vehicles (UAVs) - in this study could be employed to figure out the most appropriate
will be the addendum. This is because, as indicated in the overall phase/place within the operational value chain to apply any one
ranking (shown in Fig. 6), UAVs are the next preferred option, and the technology.
investment and operational cost for both technology options ranges
from 50,220 to 57,780. 5.5. Managerial implications

5.4. Limitation of study This study can particularly help managerial decision makers because
it has;
This study was particularly limited due to its hypothetical nature
which involved just one decision maker. However, a typical mine would a) Provided an insightful and clearer understanding of how various new
have multiple decision makers. Therefore, further research can investi­ and emerging technologies can be adopted along the existing mining
gate a technology decision problem modelling more decision-makers. In value chain;
addition, other MCDM techniques can be tested and the computational b) Provided the basis for which a technology business case, strategy or
results compared to that of this study. Also, for a further research, a roadmap can be built for any technology adopting mining company;
reverse approach can also be considered, such that the method proposed and

13
O. Dayo-Olupona et al. Resources Policy 69 (2020) 101879

Fig. 9. Original Criteria Weightage used in this Study and the Results.

Fig. 10. Results of the analysis after setting equal criteria weightage.

c) Provided a stepwise methodological approach that has the ability to preferred technology alternative if a single choice of technology is
solve decision problems with both fuzzy and precision-based (non- needed for adoption. However, with the budget made available by the
fuzzy) characteristics, making it ideally suitable for real life mine mining company, more than one technology can be chosen. Therefore,
decision-making scenarios. This is because, most times, while the best ranked two, which are AI (A3) and UAVs (A4), can be adopted
selecting technologies or evaluating them for a proposed project, the by the company.
information available to make those decisions are usually imprecise The hierarchal weight calculated using the AHP techniques revealed
and uncertain. that the “Economic” criteria is the main criterion because it has the
highest weight relative to others. The “Operational Safety” criterion was
6. Conclusion also the next highest criterion. This reveals that for the subsequent
evolution process, the decision maker payed more attention to the
This study has demonstrated the application of two MCDA methods “Economic” and “Operational Safety” criteria in comparison to the
in evaluating emerging technologies for the selection of a single or group others.
of preferred technologies for any adopting mining company. The It is important to note that the concept presented in this study,
structure of the decision problem resolved by the MCDA method con­ provided the fundamentals for more specific and targeted studies in the
sisted of six technology alternatives – AR, AA, AI, UAVs, IoEs and IoTs – space of technology and its adoption within the metals, mining and
from a cumulative of about 550 that were assessed against four criteria. mineral processing industry. It can also play a critical part in making the
This evaluation procedure used the AHP method to determine the hi­ managerial technological decision making easier.
erarchal weight of each decision-making criterion and its consistency
while, the PROMETHEE method was used to carry out the overall pro­ Acknowledgements
cess evaluation. Additionally, the fuzzy set theory was however infused
into the hierarchical structure analysis to evaluate the quantitative The work reported in this paper is part of an M. Sc. Research report in
economic criterion and also in a quantitative manner to curb uncertainty the School of Mining Engineering at the University of the Witwa­
and imprecision. tersrand, Johannesburg, South Africa.
The result of the overall evaluation shows that AI (A3) is the most

14
O. Dayo-Olupona et al. Resources Policy 69 (2020) 101879

Appendix

Fuzzy Data Calculation

For technology A1, at the end of year 0

G10 = (0, 0, 0),

C10 = (0, 0, 0),

D10 = (0, 0, 0),

K10 = (9300.0, 10000.0, 10700.0)

(1 − T1 )G10 = (0, 0, 0)

(1 − T1 )C10 = (0, 0, 0)

D10 T1 = (0, 0, 0)
By Equation (Eq.) 7
X10 = (− 10700.0, − 10000.0, − 9300.0)
For technology A1, at the end of year 1
G11 = (7905.0, 8500.0, 9095.0),

C11 = (860.0, 2000.0, 2140.0),

D11 = (3100.0, 3333.3, 3566.7),

K11 = (0, 0, 0)

(1 − T1 )G11 = (4743.0, 5100.0, 5457.0)

(1 − T1 )C11 = (1116.0, 1200.0, 1284.0)

D11 T1 = (1240.0, 1333.3, 1426.7)

By Equation (Eq.) 7
X11 = (7905.0, 8500.0, 9095.0)(1 − 0.4) − (1860.0, 2000.0, 2140.0)(1 − 0.4) + (3100.0, 3333.3, 3566.7)(0.4)

X11 = (4699.0, 5233.3, 5767.7)

The cash flow for the two subsequent years – Year 2 and Year 3 – will have the same values as Year 1
(4699.0, 5233.3, 5767.7) (4699.0, 5233.3, 5767.7) (4699.0, 5233.3, 5767.7)
NPV1 = (− 10700.0, − 10000.0, − 9300.0) + + +
(1.12, 1.12, 1.12)1 (1.12, 1.12, 1.12)2 (1.12, 1.12, 1.12)3

NPV1 = (586.21, 2569.58, 4552.96)

Using (Eq.) 8.
This same procedure was carried out for each of the technology option.

References Deloitte, 2015. Tracking the trends 2015: the top 10 issues mining companies will face
this year. Tech. rep., Deloitte. Retrieved from. https://www2.deloitte.com/content
/dam/Deloitte/fpc/Documents/secteurs/energie-et-ressources/deloitte_etude-trac
Abdullah, L., Chan, W., Afshari, A., 2019. Application of PROMETHEE method for green
king-the-trends-2015-en.pdf.
supplier selection: a comparative result based on preference functions. J. Indus. Eng.
Deloitte, 2019. Tracking the trends 2019: the top 10 issues transforming the future of
Int. 15 (2), 271–285.
mining. Tech. rep., Deloitte. Retrieved from. https://www2.deloitte.com/content
Amadi-Echendu, J., Lephauphau, O., Maswanganyi, M., Mkhize, M., 2019. Case studies
/dam/Deloitte/global/Documents/Energy-and-Resources/gx-er-tracking-the-trends-
of technology roadmapping in mining. J. Eng. Technol. Manag. 23–32.
2019.pdf. Accessed: 31 May 2019.
Bartos, P.J., 2007. Is mining a high-tech industry? : investigations into innovation and
Dickerson, A.W., 1997. Utilising New Technology. The Australasian Institute of Mining
productivity advance. Resour. Pol. 32 (4), 149–158.
and Metallurgy, Sydney, Australia.
Bryant, P., 2015. The case for innovation in the mining industry [Online] Available at:
Drake, J.I., Hart, J.C., Monleón, C., Toro, W., Valentim, J., 2017. Utilization of multiple-
http://www.ceecthefuture.org/wp-content/uploads/2016/01/Clareo_Case-for-Inn
criteria decision analysis (MCDA) to support healthcare decision-making FIFARMA,
ovation-in-Mining_20150910_lo.pdf. Accessed 13 May 2019.
2016. J. Market Acc. Health Policy 5 (1), 10.
Chan, F., Chan, M., Tang, N., 2000. Evaluation methodologies for technology selection.
Durrant-Whyte, H., Geraghty, R., Pujol, F., Sellschop, R., 2015. How Digital Innovation
J. Mater. Process. Technol. 107, 330–337.
Can Improve Mining Productivity. McKinsey. Tech. rep., McKinsey Inc.. Retrieved
Crawford, A., 2018. Innovation in Mining: Report to the 2018 International Mines
from. https://www.mckinsey.com/industries/metals-and-mining/our-insights/ho
Ministers Summit. Technical Report. The International Institute for Sustainable
w-digital-innovation-can-improve-mining-productivity. Accessed: 24 Jan 2019.
Development, Winnipeg (IISD).
Ernst & Young, 2018. Worldwide Capital and Fixed Asset Guide, Report. Ernst & Young
Dey, P.K., Ramcharan, E.K., 2008. Analytic hierarchy process helps select site for
Global Limited, United Kingdom.
limestone quarry expansion in Barbados. J. Environ. Manag. 88, 1384–1395.

15
O. Dayo-Olupona et al. Resources Policy 69 (2020) 101879

Farrelly, C.T., Malherbe, G., Gonzalez, J., Bassan, J., Franklin, D.C., 2012. The network Runge, I.C., 1995. Economics, Technological Change and the Knowledge Problem. The
centric mine. In: International Mine Management Conference. The Australasian Australasian Institute of Mining and Metallurgy, Brisbane, p. 6.
Institute of Mining and Metallurgy, Melbourne, Australia. Saaty, T.L., 1995. Transport planning with multiple criteria: the analytic hierarchy
Goumas, M., Lygerou, V., 2000. An extension of the PROMETHEE method for decision process applications and progress review. J. Adv. Transport. 29 (1), 81–126.
making in fuzzy environment: ranking of alternative energy exploitation projects. Sabaei, D., Erkoyuncu, J., Roy, R., 2015. A review of multi-criteria decision making
Eur. J. Oper. Res. 123 (3), 606–613. methods for enhanced maintenance delivery. Procedia CIRP 37, 30–35.
Haddad, M., Sanders, D., 2018. Selection of discrete multiple criteria decision making Samimi, N.F., Shahriar, K., Basçetin, A., Ghodsypour, S.H., 2009. Practical applications
methods in the presence of risk and uncertainty. Oper. Res. Perspect. 5, 357–370. from decision-making techniques for selection of suitable mining method in Iran.
Hudej, M., Vujic, S., Radosavlevic, M., Ilic, S., 2013. Multi-variable selection of the main Gospodarka Surowcami Mineralnymi - Min. Res. Manag. 25 (3), 57–77.
mine shaft location. J. Min. Sci. 49 (6), 950–954. Savic, M., Nikolic, D., Mihajlovic, I., Zivkovic, Z., Bojanov, B., Djordjevic, P., 2015.
Hwang, C.L., Yoon, K., 1981. Multiple Attribute Decision Making. Springer Berlin Multi-criteria decision support system for optimal blending process in zinc
Heidelberg, Berlin, Heidelberg. production. Miner. Process. Extr. Metall. Rev. 36 (4), 267–280.
Jacobs, J., 2016. Creating a Technology Map to Facilitate the Process of Modernisation Sitorus, F., Cilliers, J.J., Brito-Parada, P.R., 2019. Multi-criteria decision making for the
throughout the Mining Cycle. Dissertation. University of Pretoria, Pretoria. choice problem in mining and mineral processing: applications and trends. Expert
Retrieved from. http://hdl.handle.net/2263/61296. Syst. Appl. 121, 393–417.
Jang, H., Topal, E., 2020. Transformation of the Australian mining industry and future Stojanović, C.P., Bogdanović, D., Urosević, S., 2015. Selection of the optimal technology
prospects. Min. Technol. 1–15. for surface mining by multi-criteria analysis. Kuwait J. Sci. 42 (3), 170–190.
Jordaan, J.T., Hendricks, C., 2009. The Challenge of Technology Adoption and Taha, Z., Banakar, Z., Tahriri, F., 2011. Analytical hierarchy process for the selection of
Utilisation in the Mining Industry – A Focus on Open Pit Mining Technologies. The advanced manufacturing technology in an aircraft industry. Int. J. Appl. Decis. Sci. 4
Southern African Institute of Mining and Metallurgy, pp. 69–86. (2), 148–154.
Koksalan, M., Wallenius, J., Zionts, S., 2011. Multiple Criteria Decision Making: from Talukder, B., Hipel, K.W., 2018. The PROMETHEE framework for comparing the
Early History to the 21st Century. World Scientific Publishing Co Pte Ltd, sustainability of agricultural systems. Resources 7 (4), 74–96.
Hackensack, NJ. Thomaidis, N.S., Nikitakos, N., Dounias, G.D., 2006. The evaluation of information
Komolavanij, S., 1995. Capital Budgeting in a Fuzzy World. The University of Texas at technology projects: a fuzzy multicriteria decision-making approach. Int. J. Inf.
Arlington, Texas, United States. Technol. Decis. Making 5 (1), 89–122.
Musingwini, C., Minnitt, R.C.A., 2008. Ranking the Efficiency of Selected Platinum Mining Triantaphyllou, E., 2000. Multi-criteria Decision Making Methods: A Comparative Study.
Methods Using the Analytic Hierarchy Process (AHP). South Africa, Third International Kluwer Academic Publishers, Dordrecht, The Netherlands.
Platinum Conference ‘Platinum in Transformation’. The Southern African Institute of Vujic, S., Hudej, M., Miljanovic, I., 2013. Results of the promethee method application in
Mining and Metallurgy, Sun City, South Africa. selecting the technological system at the majdan III open pit mine. Arch. Min. Sci.
Ordoobadi, S.M., 2012. Application of ANP methodology in evaluation of advanced 58, 213–219.
technologies. J. Manuf. Technol.Manag. Bradford 23 (2), 229–252. Wang, C., Tu, S., 2015. Selection of an appropriate mechanized mining technical process
Owusu-Mensah, F., Musingwini, C., 2011. Evaluation of ore transport options from Kwesi for thin coal seam mining. Math. Probl Eng. 2015 (893232), 10.
Mensah Shaft to the mill at the Obuasi mine. Int. J. Min. Reclamat. Environ. 25, World Economic Forum (WEF), 2017. Digital transformation initiative: mining and
109–125. metals industry. [Online]. Available from: http://reports.weforum.org/digital-tr
Poulin, P., Austen, L., Scott, C.M., Waddell, C.D., Dixon, E., Poulin, M., Lafrenière, R., ansformation/wp-content/blogs.dir/94/mp/files/pages/files/wef-dti-mining-and-
2013. Multi-criteria development and incorporation into decision tools for health metals-white-paper.pdf, 36. Accessed: 31 May 2019.
technology adoption. J. Health Org. Manag. Bradford 27 (2), 246–265. Yuen, K.K.F., Ting, T.O., 2012. Textbook selection using fuzzy PROMETHEE II method.
Readett, D., 2006. Selecting and Implementing the Right Technology. The Australasian Int. J. Future Comp. Commun. 1 (1), 76–78.
Institute of Mining and Metallurgy, Kalgoorlie, Western Australia. Zavadskas, E., Turskis, Z., Kildiene, S., 2014. State of art surveys of overviews on MCDM/
MADM methods. Technol. Econ. Dev. Econ. 20, 165–179.

16