INTRODUCTION
Personnel selection is the process of choosing individuals who match the qualifications
required to perform a defined job in the best way. It determines the input quality
of personnel and plays a important role in human resource management. Increasing
competition in global markets urges organizations to put more emphasis on personnel
selection process. Important issues such as changes in organizations, work,
society, regulations and marketing have an influence on personnel selection
and recruiting. Organizations differ with respect to the procedures and budgets
for recruiting, selecting and orienting people (Karsak, 2001).
Some firms make a strategic decision to choose the best candidate by utilizing
rigorous and costly selection procedures, while others decide to fill positions
quickly and inexpensively based only on the information stated on the application
forms. Nonetheless, the growing importance attached to personnel selection process
has paved the way for analytical decision making approaches.
Organizations today are making abundant changes internally to cope with a highly
turbulent external environment. Frequent reorganizing, downsizing, rightsizing,
hierarchical flattening, teaming and outsourcing shape the selection process;
which is influenced by the fact that many people are experiencing major difficulties
in their attempts to adapt to the uncertainties of career life (Brousseau
et al., 1996). In general, human resource practices and climate have
considerable impact on how the shock of downsizing ultimately translates to
organizational performance (Trevor and Nyberg, 2008).
Many studies have reported a positive association between various human resources
practices and objective and perceptual measures of selecting human resources,
some authors have expressed concern that results may be biased because of methodological
problems (Kulik et al., 2007; Le
et al., 2007). Traditional methods for selection of human resources
are mostly based on statistical analyses of test scores that are treated as
accurate reflections of reality. Modern approaches, however, recognize that
selection is a complex process that involves a significant amount of vagueness
and subjectivity (Kulik et al., 2007).
In general, personnel selection, depending on the firm’s specific targets, the availability of means and the individual preferences of the Decision Makers (DMs), is a highly complex problem. The multicriteria nature of the problem makes MultiCriteria Decision Making (MCDM) methods and cope with this, given that they consider many criteria at the same time, with various weights and thresholds, having the potential to reflect at a very satisfactory degree the vaguemost of the timespreferences of the DMs. In this study, ELECTRE method is suggested to solve personnel selection problem using multicriteria decisionmaking process.
In most of the situations where a decision must be taken, it is rare for the
DM to have in mind a single clear criterion (Figueira et
al., 2005). Such situations, where a singlecriterion approach falls
short, refer to as MCDM problems. Many terminologies have been proposed for
the categorization of MCDM problems. The dominant terms are the one of MultiCriteria
Decision Analysis (MCDA) or MultiAttribute Decision Making (MADM), for problems
in which the DM must choose from a finite number of explicitly available alternatives
characterized by a set of multiple attributes (or criteria) and the one of MultiObjective
Mathematical Programming (MOMP) or MultiObjective Decision Making (MODM) that
deal with decision problems characterized by multiple and conflicting objective
functions that are to be optimized over a feasible set of decisions. Here, the
alternatives are not explicitly known a priori (Figueira et
al., 2005). In what follows, the main categories of MCDM are presented.
One class of approaches that deal with subjectivity includes techniques based
on the wellknown Analytic Hierarchy Process (AHP) which reduces complex decisions
to a series of pair wise comparisons and synthesizes the results. AHP and its
extensions have been utilized extensively in the selection of human resources.
Typical applications include the ones presented by Lai (1995),
Iwamura and Lin (1998) and Labib
et al. (1998). Albayrak and Erensal (2004)
used AHP, which determines the global priority weights for different management
alternatives, to improve human resource performance outcomes. A detailed review
of various applications of AHP in different settings is provided by Vaidya
and Kumar (2006). Lai (1995) describes the employee
selection process as a multiobjective decisionmaking problem. Iwamura
and Lin (1998) explain that the employee selection process requires the
accomplishment and aggregation of different factors. Labib
et al. (1998) suggested an employee selection process that uses the
AHP and has four stages.
The other contemporary methods in the employee evaluation and selection are
artificial intelligence techniques that are the fuzzy sets and neural networks.
In contrast to conventional sets where a given value v is either included or
not included in a set A, in fuzzy set theory each value is associated with a
certain grade of membership in set A. This grade is expressed by a membership
function that reflects the degree to which it can be argued that value v is
included in A. Examples of such approaches can be found by Laing
and Wang (1992), Yaakob and Kawata (1999), Lovrich
(2000) and Wang et al. (2006). Lazarevic
(2001) introduced a twolevel fuzzy model for minimizing subjective judgment
in the process of identifying the right person for a position. And Royes
et al. (2003) proposed a combination of fuzzy sets and multicriteria
tools for employee selection. In a similar approach, Golec
and Kahya (2007) proposed a hierarchical structure and use a fuzzy model
that has two levels: evaluation and selection. The first level employs a heuristic
algorithm which evaluates candidates according to measure indicators whereas
the second level selects the candidate using a fuzzy rulebased approach.
Some studies focused on proposed Expert Systems (ESs) or decision support systems
to assist personnel selection. Wabalickis (1988) studied
the capability of ES and pointed out that it has the potential to assist with
tasks for selecting new employees, matching people with jobs, training new and
old employees and so on. Later, a working ES named EXPER (Suh
et al., 1993) was developed to assist managers in making job placement
decisions, where employees were evaluated with respect to test scores, performance
ratings, aptitude scores and so on and then were matched with specific jobs
within an organization. Brunsson et al. (1998)
developed and tested a rulebased ES, BOARDEX, to perform the Yes/No vote to
screen officer personnel records in the first phase of board procedure. Experiment
on a mock officer personnel records showed that BOARDEX was successful at selecting
the records. Drigas et al. (2004) presented an
expert system using neurofuzzy techniques that investigate a corporate database
of unemployed and enterprises profile data for evaluation of the unemployed
at certain job position. This study uses a sugeno type neuro fuzzy inferences
system for matching an unemployed with a job position. Huang
and Chen (2005) proposed a data mining framework based on decision tree
and association rules to generate the useful rules for personnel selection.
The useful rules were extracted from the relationships between personnel profile
data and their work behaviors. Finally, 30 meaningful rules were chosen to develop
the recruitment strategies.
MATERIALS AND METHODS
MultiCriteria DecisionMaking (MCDM) is one of the most widely used decision
methodologies in the sciences, business, government and engineering worlds.
MCDM methods can help to improve the quality of decisions by making the decisionmaking
process more explicit, rational and efficient (Wanga and
Triantaphylloub, 2008). Some applications of MCDM in engineering include
the use on flexible manufacturing systems (Wabalickis, 1988),
layout design (Cambron and Evans, 1991), integrated manufacturing
systems (Putrus, 1990) and the evaluation of technology
investment decisions (Boucher and Mcstravic, 1991).
Many methods have been proposed to analyze the data of a decision matrix and
rank the alternatives. Often time’s different MCDM methods may yield different
answers to exactly the same problem (Triantaphyllou, 2000)
and also some of the methods use additive formulas to compute the final priorities
of the alternatives.
The ELECTRE evaluation method is widely recognized for highperformance policy
analysis involving both qualitative and quantitative criteria. However, a critical
advantage of this evaluation method is its capacity to point the exact needs
of a decision maker and suggest an appropriate evaluation approach. The discordance
indices of modified ELECTRE evaluation method are used to explain the significance
of modified evaluation standards (Huang and Chen, 2005).
The ELECTRE method is a well known method, especially in Europe too. It has
been widely used in civil and environmental engineering (Hobbs
and Meier, 2000). Applications include the assessment of complex civil engineering
projects, selection of highway designs, site selection for the disposal of nuclear
waste, water resources planning (Anand Raj, 1995) and
waste water (Rogers et al., 1999) or solid waste
management (Hokkanen and Salminen, 1997) etc.
ELECTRE was conceived by Roy (1991) in response to
deficiencies of existing decision making solution methods. ELECTRE is more than
just a solution method; it is a Philosophy of decision aid the philosophy is
discussed at length by Roy (1991). However, for this
study we specifically concentrate on what is referred to as ELECTRE. ELECTRE
has evolved through a number of versions (I, II, II, IV, V, IS, A); all are
based on the same fundamental concepts but are operationally somewhat different
(Huang and Chen, 2005).
Two separated phases are designed in order to address the research methodology,
the stages are shown in Fig. 1 and are presented as follows:
• 
Phase 1: The first phase of this study is designed
in order to select and consider suitable criteria and personnel in one of
a sector of Telecommunication’s Company respectively. The way of data
collection that is applied for this phase is questionnaire. By using Comparison
Matrix with one part of collected data that have been prepared by experts,
the weights of criteria will be computed. After computing weights of criteria,
specifying of Consistency will be executed too. If the Consistency is less
than 0.1, then we use ELECTRE method for preranking personnel. This phase
is especially important because it provides the knowledge platform and preselecting
personnel for next phase 
• 
Phase 2: The applied methodology for this phase is
based on output of previous phase and the method used is AHP. In this phase,
after identifying the level of personnel, we apply AHP method when at least
one of personnel’s grades was placed in the same with another. In this
way, specifying of Consistency will be executed too. In both of phases,
if Consistency of data is less than 0.1, revision of pairwise comparison
must be done. At the end of this phase, all of personnel which had been
considered will be sorted in different level 

Fig. 1: 
Research framework 
After specifying relative criteria and also considering five people as alternatives,
computing the weights of criteria were started by using comparison matrix.
Data was gathered from five expert’s point of view in one of sector in Telecommunication Company. Following steps will be shown the way of solving an application problem in ELECTRE method and finally with AHP method it will rank the result of ELECTRE that some personnel were in the same level.
Steps of ELECTRE method: Asgharpour (2008):
• 
Step 1: Calculate the normalized decision matrix 
• 
Step 2: Calculate the weighted normalized decision
matrix 
We assumed that W is a diagonal matrix (nxn) which values of its main diameter are W and the rest values are zero.
• 
Step 3: Determine the concordance and discordance set 
When this condition is true then we put 1 in its place otherwise we put 0.
We will also apply for discordance set as followed:
It is obvious that S_{kl} and D_{kl} are opposite
then places of 0 belong to.
• 
Step 4: Calculate the concordance matrix 
In this matrix (I) is {k,l = 1,2,3...m, k≠1}, so each element of matrix includes sum of element(s) W, that they depend to S_{kl} .
Therefore, each elements of S_{kl} will be between: 0≤Ik,l≤1.
• 
Step 5: Calculate the discordance matrix 
During computing matrix of NI, it is necessary that {k,l = 1,2,3...m,
k≠1}, so each elements of matrix will be computed as follow:
in nominator and denominator respectively
• 
Step 6: Determine the concordance dominance matrix 
Dimension of matrix F and matrix I (in step 4) are the same but for finding
matrix F, it is needed to compute threshold amount ()
as follow:
(m is dimension of matrix).
Matrix F can be calculated by using matrix I if each corresponding elements
of matrix I, are divided to (Threshold
amount of this step).
The above inequalities mean that if each element of matrix I, is greater than
or equal to ,
then 1 would be set in matrix F (corresponding element).
• 
Step 7: Determine the discordance dominance matrix.So
we calculate matrix of G 
Matrix G can be calculated by using matrix NI, if each corresponding elements
of matrix NI, are divided to (Threshold
amount of this step).
Also, the above inequalities mean that if each element of matrix NI, is less
than or equal to ,
then 1 would be set in matrix G (corresponding element).
• 
Step 8: Determine the aggregate dominance matrix 
We also compute matrix H. P is means personnel.
So, matrix H is performed by multiplying corresponding elements of F and G.
• 
Step 9: Eliminate the less favorable alternative and
rank them 
Finally, we must scan the columns of matrix H, each column that has the least
amount of number 1 should be chosen as the best one.
Analytic Hierarchy Process (AHP) is widely used multi criteria decision making
method introduced by Satty (1980) and it resolves decision
making problems by structuring each problem into a hierarchy with different
levels of criteria. In other word, AHP structures a decision problem into a
hierarchy and evaluate multi criteria tangible and intangible factors systematically.
AHP also has been applied in numerous fields (Forman and
Gass, 2001; Vargas, 1990; Zahedi,
1986) including many software selection decisions. The AHP method involves
four steps to solve a decision problem (Zahedi, 1986;
Lin and Yang, 1996; Tam and Tummala,
2001):
Steps of AHP method:
• 
Step 1: Structuring the decision problem 
Structure the hierarchy from the top (goal) through the intermediate levels
(criteria, subsequent levels depend on) to the lowest level which usually contains
the list of alternatives.
• 
Step 2: Creating pair wise comparison matrix 
After constructing AHP model, the priorities should be done. Weights are assigned
to each criterion and sub criterion. These weights are assigned through a process
of pair wise comparison. In pair wise comparison, each objective is compared
at a peer level in terms of importance. In this time, a set of pairwise comparison
matrices (size nxn) for each of the lower levels with one matrix for each element
in the level immediately above by using the relative scale measurement shown
in Table 1 is constructed. The pairwise comparisons are done
in terms of which element dominates the other.
• 
Step 3: Determining normalized weights 
So, by using each pairwise comparison Matrices, weight of each row was computed
by matrix of W.
• 
Step 4: Synthesize the priorities 
The final step is to synthesize the solution for the decision problem in order
to obtain the set of priorities for alternatives. After computing the weight
of alternatives in respect to sub criteria and then sub criteria in respect
to criteria and also criteria in respect to goal from step 3 (in the level immediately
above), they are aggregated to produce composite weights which used to evaluate
decision alternatives.
If each comparison matrix was filled randomly, it will be shown by the Consistency
Ratio. The Consistency Ratio (CR) is an indicator which mathematically approximates
level of pairwise comparisons. It work as a function of maximum Eigen value
and size of the matrix (Consistency Index), which is then compared to similar
values if the pairwise comparison had been merely random (Random Index). The
consistency of a comparison is totally acceptable for pragmatic purposes if
it is not greater than 0.1 (Satty, 1980).
Table 1: 
Determining the weights of criteria by comparison matrix 

W= {0.264, 0.234, 0.075, 0.117, 0.053, 0.175, 0.082} 
Table 2: 
The normalized decision matrix 

Table 3: 
The weighted normalized decision matrix 

Table 4: 
Stages of concordance and discordance set 

Numeric example: In Feb. 2009, by using seven criteria, one sector of
Telecommunication Company in KhorasanIran, must be chosen one of the five people
which have passed the exam. Here are criteria that have been mentioned above.
C1 
= 
Ability to work in different business units 
C2 
= 
Past experience 
C3 
= 
Team player 
C4 
= 
Fluency in a foreign language 
C5 
= 
Strategic thinking 
C6 
= 
Oral communication skills 
C7 
= 
Computer skills 
Calculating the weights of criteria has been computed by using comparison matrix.
Meanwhile, data was gathered from five expert’s point of view in one of
sector in Telecommunication Company as shown in Table 1. For
solving this kind of problem, First of all by calculating the normalized matrix
in Table 2, we will prepare the matrix for weighted normalized
matrix (Table 3) by applying V_{ij} = N_{ij}xW_{ij}.
Table 5: 
Concordance matrix 

Table 6: 
Discordance matrix 

After that, for computing concordance and discordance matrix in Table
5 and 6, respectively, we must calculate concordance and
discordance set as shown Table 4. We will also compute concordance
(Table 7) and discordance (Table 8) dominance
matrix, by using Table 5 and 6 respectively.
Finally, the result will be extract from Table 9.
Table 7: 
Concordance dominance matrix 

Table 8: 
Discordance dominance matrix 

Table 9: 
Aggregate dominance matrix 

And finally, we can eliminate the less favorable alternative and rank them. In ELECTRE method, the best personnel will be P3 and P2 (in equal value) and they were followed by P5, P1 and P4. By using AHP, we solve this problem and determined that P3 will be preferred to P2. So, the result is: P3>>P2>>P5>>P1>>P4
In many times ELECTRE method cannot sort alternatives in different rank, so in this method, authors have applied from hybrid ELECTRE with AHP methods in order to solve this kind of problems. This methodology for personnel selection has not been seen in previous findings so far.
CONCLUSIONS
In this study, we presented a MCDM methodology for selecting employees to cover organizational positions. The method was applied using data from a real case in the Telecommunication sector of Iran. To increase the efficiency and easeofuse of the proposed model, simple software such as MS Excel can be used. Evaluation of the candidates on the basis of the criteria only will be sufficient for the future applications of the model and implementation of this evaluation via simple software will speed up the process.
The limitation of this article is that ELECTRE ignores the fuzziness of executives’ judgment during the decisionmaking process. Besides, some criteria could have a qualitative structure or have an uncertain structure which cannot be measured precisely. In such cases, fuzzy numbers can be used to obtain the evaluation matrix and the proposed model can be enlarged by using fuzzy numbers. For the future research, the authors suggest the other multicriteria approaches such as ELECTRE III and fuzzy outranking methods to be used and to be compared in justification of the personnel selection problem.
Finally, ELECTRE may be employed to address several human resource issues other
than the selection process. Typical applications include the evaluation of training
and development programmes and the assessment of individual employees or work
groups. The method may also be applied in other business problems, not directly
related to human resources. Examples of such applications include project selection
and supplier selection in a supply chain.