Comparison of Job Position Based Promotion Using: VIKOR, ELECTRE And Promethee Method

The long-term prospect of the company's progress is determined by the quality of human resources (HR), The urgency to maintain the company's survival it takes a reliable and futuristic leader. Measuring tool that can be used is none other than the performance of human resources. Of course with potential leaders will provide the vision of the company's mission to grow and expand. Leadership selection process can be done with promotion positions based on performance preference. The purpose of this research is to conduct selection of performance promotion based on performance using Multi-Criteria Decision Making (MCDM) selection methods such as Analytic Hiererachy Process (AHP), VIKOR, ELECTRE, and Promethee, in addition to proving the result of a number of methods based on MCDM such as AHP for the determination of preferences data design, while for data analysis using three methods that will be compared the results of VIKOR, ELECTREE and Promethee. The measurable performance bases for promotional positions are viewed from seven criteria described in the modeling hierarchy using the concept of analytic hierararchy process in the discussion of this paper. The results obtained provide an interpretation that the promethee method is closer to the actual results, while the vikor method is almost close to the results of truth and far away when compared with the electre method. Thus, it can be concluded that the best method for placing employee positions is promethee method.


I. INTRODUCTION
The quality of human resources is paramount to the progress of the company, because individualized expertise can be measured from individual performance [1], this is a very important thing as a benchmark in selecting a potential leader through the promotion process. Promotion can be used as a way to find employees who have good achievements to be able to lead companies in advancing and developing the company's wings in running the company's business. [1].
The purpose of this study is to compare the results of several methods to provide the best solution for achieving the objectives of the company and to compare the results of the promotion by using three methods, namely VIKOR, ELECTRE, and Promethee, how decisions are generated using the three methods . It is very difficult to do a combination of a number of methods, we must first know the function of the method to be used and included in the group how the method can be used well, as well as methods that include the category of multicriteria decision making that can be collaborated with one another and can be used to find promoted employees [2]. Thus, certain methods are needed that can be used well to carry out promotions for employees who have high achievements and deserve to lead the company, for this reason the analytic hierarchy process is very supportive for the selection process for promotion [3]. In this study, collaboration between multi-criteria decision making will be carried out which is able to find the best solution in the best promotion promotion process for employees, where the three methods support each other to provide the best results in the selection process. It is indeed quite different from VIKOR which merely uses the concept of index in determining priorities, far away from being mixed with ELECTRE and Promethee [4] in addition to MCDM.
MCDM is a multi-criteria technique that has its own advantages in determining priorities or ranking in the selection process, one of which is the use of iterations to get the eigenvector value and eigenvector values can be used as input preferences for other multi-criteria methods such as VIKOR, ELECTRE, and Promethee, This is a ranking in the implementation of AHP. To use AHP, there are three levels including the Objectives, Criteria, and alternatives. This aims to simplify very complex problems to be as simple as possible. [6], [7], [8]. Thus the use of the VIKOR method is better known for multi-criteria decision making [9], As previously discussed, VIKOR is indeed the best spotlight that is widely used in index ranking [10], and other multi-criteria of ELECTRE and Promethee.
To provide an assessment related to performance, of course, determine the criteria set out in the employee selection process, for this study, of course, must understand more about the use of a number of criteria. Criteria related to more time tend to be understood with the smallest term is the best value of the criteria, or conversely the largest value is the best value of the criteria [10]. The ranking method with modeling aims to simplify the problem into the simplest condition of the problem, so it is very important to determine the criteria for obtaining optimal decision results by conducting an appropriate evaluation of a number of criteria that are set to pass the selection stage [11].
II. METHOD Employee job preferences in companies that are not mentioned determine the progress of the company, so that it requires a number of appropriate criteria to choose the best leader, this assessment is determined by company policy which is determined to be seven assessment criteria, by seeing competition outside these criteria to be a company decision to remain a company who is able to lead the market.
The collecting data, a leader has done a number of samples from employees with seven criteria that have been set and have been done by experts to give value to a number of employees. This has been done for several periods, then the average value of each employee is taken.
The AHP method in this research is used as a key to provide strong input on the preferences used with the concept of elimination which can be applied significantly by the VIKOR, ELECTRE, and Promethee methods and compared with one another. The results can be seen and conclude which of the elimination methods is the best in making elimination decisions.

Promotion of Position.
To determine promotion, it is necessary to pay attention to the determination of a number of criteria that greatly determine the success of the position process. Criteria are carefully selected and must have the weight of each criterion that is used as a benchmark to determine the selected employee, while performance is an effort that has been carried out by the figure of an employee who has carried out his obligations in carrying out the standard of work he has carried out [3], [12], [13].
The criteria used for promotion include seven criteria: (1) intelligence, (2) planning, (3) dependence, (4) reaction behavior, (5) failed jobs, (6) quantity of work, and (7) knowledge of job. Each criterion has two special rating categories, rating with appreciation value meaning that the score is the best value or depreciation value the lowest value is the best. For category of appreciation value criteria are planning, quantity of work, and knowledge of job, while those belonging to depreciation value criteria category are imagination dependence, reaction behavior, and failed of job. Each preference gives unequal meanings in its usage, the amount of preference can be interpreted as the best MAX value or it can be interpreted that the MIN value is the best.

Analytic Hierarchy Process (AHP).
To formulate alternatives in an analytic hierarchy process, of course, must be in the same level and there is no multilevel term, this is done to provide a valid alternative and there is no aspect of jealousy from alternative availability, so it must be done very carefully and can be accounted pairwise . So that the decision making can be understood by the office holders and can be accepted in a generous manner [14]. AHP can be used in real life in everyday life, because it can be implemented logically and the scale of comparison [15], and can be applied in the form of quantitative research and qualitative research, or a combination of both [16] which can be used to solve difficult problems even very difficult ones can be easily simplified [17] there are so many considerations that must be included in determining the process of value of interest, because numerical assessment greatly determines the quality of decisions, especially in subjectivity assessments that are very concentrated in providing value comparisons for ranking criteria and alternative levels. [2], [18]. The modeling hierarchy is very decisive in decision making as a preference that can be used in various fields of selection and evaluation, to solve this problem can be done by compiling a matrix as a comparison of algebra matrix pairs.

Multi-Criteria Decision Making (MCDM).
The MCDM method is indeed well known to a number of researchers especially in decision making with good algorithms by considering carefully and easily understood [20].
The MCDM method uses weighting concepts from each criterion and alternative that are based on clear algorithms and many variants of this method. [20].
Multi criteria technique known in MCDM is one of the best ways that can be applied and is quite easy to understand to formulate both criteria and alternatives to find agreement and unity of the resulting understanding of decision support [21].
The MCDM method has been widely recognized for its reliability in supporting decision making implemented in the AHP method and there are many variations that can be developed in AHP [5], [10], [22], [23], [24]. MCDM is a method that can be used to rank a number of criteria by setting alternatives that pay attention to similarities in the level. Comparison between them will get the value set as priority to be reused as a preference that leaves the ideal closeness and is used as a joint function in the VIKOR, ELECTRE, and Promethee methods. In the VIKOR method the combined function used aims to eliminate the criterion function unit, so that the vikor is included in one of the elimination methods besides ELECTRE and Promethee [5], So that it can be seen the limit of the value produced by the AHP method, by paying attention to the amount of eigenvector value produced through several iterative stages in MCDM to obtain an optimal eigenvector value for a number of criteria (Table 1 and Table  2).  [25] Each matrix element in (table 1) is in the position of certain rows and certain columns, rows can be represented with dimensions i and columns can be represented by dimensions j, each matrix element with the concept of algebra matrix has a reversal value where columns turn into rows and rows changes to a column for the input element value in the lower triangle matrix [17], [25]. The usefulness of MCDM as outlined in AHP is the best way to get preference values for processing data later and to eliminate dependency functions from the next method to be used, in this case VIKOR, ELECTRE, and Promethee. Multi-criteria with AHP method provides optimal results in setting values against eigenvector, eigenvector value is said to be optimal because it appears there is no difference between the last iteration and the previous iteration or can be said to be worth zero of decimal.

VIKOR.
VIKOR stands for (Vise Kriterijumska Optimizacija I Kompromisno Resenje) and is one of the elimination methods used to determine the rank of a number of alternatives by determining indexed values [26].
To determine the results with the VIKOR method, you must first find the ratio for each solution of positive ideal values and solutions to negative ideal values, this is done to determine the amount of the VIKOR index value to determine the smallest amount as the highest rating and vice versa the largest magnitude as the lowest [27] As the final result of the VIKOR method, the value that approaches the value of the ideal solution and VIKOR applies the concept of index taken as an alternative rating, the smallest index value. VIKOR matrix normalization is used to determine the exact location value in determining the value of the sample, performed (equation-1). After the sample R (i, j) is normalized, then the weight of the Wj criterion is multiplied by R (i, j) summed into Si shown in (equation-2) and Ri represents the largest value of each row i seen in (equation-3). (2) After knowing the weight of each row, the dimensions of Si, and S* the largest value of S, S 'the smallest value of S, R* the largest variable R value and R' the smallest variable R value, of all, can be done determination of the VIKOR Qi index, with the equation seen in (equation-4). (4)

ELECTRE
The ELECTRE method is a concept of elimination from a normalized result through a function which is further downgraded to concordance and discordance to rank. The ELECTRE method has a basic understanding of the same data as a VICOR that has been normalized before with the rules The largest value is the best value or the smallest value is the best value. Some steps must be taken to use the ELECTRE method seen in (Fig.3), MCDM combination poured into AHP and is a simplified way to be understood. Some of the equation used  So the end result can be searched by mathematical deduction of concordance and discordance resulting in a ranking of alternatives that can be decided.

Promethee.
Stages that must be known in Promethee there are some steph [28] namely: Steph 1: beginning to determine the deviation from pairwise comparisons Steph 2: Application of the preference function. Steph 3: Calculation of an overall or global preference index. Steph 4: Calculation of outranking flows the promethee-1 partial ranking (leaving flow dan entering flow). Steph 5: Calculation of outranking flows the promethee-2 as complete ranking.
Promethee has a similar calculation process with ELECTRE, in terms of determining the magnitude of the matrix normalization, see (

III. IMPLEMENTATION AND RESULTS
The description of the hierarchical model is made to clarify the problems discussed, the hierarchical model provides an overview of each level of the hierarchy consisting of three levels including (1) Goal that explains the purpose of the hierarchical model that discusses the objectives of promotion in the company, (2) Criteria that explain a number of criteria that become an objective measure that are in accordance with the objectives at (level 1), namely goals, (3) alternatives that explain the best selection of methods used for the selection process of existing alternatives. The final result of AHP is to determine the output in the form of preferences for the weight of each criterion, and the weight of the criteria produced by multi-criteria can be used as collaboration in other multicriteria methods, in this case VIKOR, ELECTRE, and Promethee. The optimal preference is generated through the AHP method and is used to rank the alternatives for a number of five selected employees based on predetermined criteria. The design of the hierarchy is made as a tool for understanding the research equipped with each level of goals, criteria and alternatives and can be seen in (fig. 3). Multi-Criteria using the AHP, used to determine the preference for the magnitude of each criterion, known under the optimal eigenvector and obtained through the iteration stage in the absence of the difference with the previous eigenvector, is the importance of the eigevector value [29]. For the value of criterion preferences, note (TABLE II). The preferences shown in (Table 2) provide an explanation of the seven criteria which are the same barometer of the criteria. The value obtained is the result of a perfect iteration of the eigenvector and as the basis for determining multi-criteria rankings elimination. Observe in (TABLE III) containing research data that will be used as master data that is able to provide a value of five employees (K1-K5) based on predetermined criteria, this master data will be used as a benchmark for each method in determining the rank of five employees as promotion, the calculation method used is VIKOR, ELECTRE, and Promethee.

VIKOR Method.
The first stage of the VIKOR method is to determine the normalization matrix, the results can be seen in (TABLE IV). Note (table iv) is the result of normalization of the master data in (table III) with the intention to determine the position weights of each criterion of five employees (K1-K5) that are promoted by the value obtained. The magnitude of the Si value describes the position of a value from the normalization results while the Ri value describes the largest value of each criterion which will be used to determine the magnitude of the vikor index for ranking decision support, the results can be seen in (TABLE VI).

TABLE VI. VALUE OF Si and Ri
After knowing the functions and uses of the Si and Ri quantities as described previously, it is time to determine the VIKOR index to determine the rank depicted in the VIKOR index quantity (Q) using (equation 4), the smallest index describing the highest ranking results and the largest index representing the lowest. The results can be seen in (TABLE VII). With reference to (TABLE III), the ELECTRE method can calculate Concordance and Discordance by using a comparison of criteria functions, note (TABLE VIII). Thus, the value of Concordance and Discordance can be simplified, The results can be seen in (TABLE IX).   (TABLE IX), the value of the concordance is the value of the upper triangle, while the discordance value in the lower triangle will be obtained by determining the ranking of the ELECTRE method, see (TABLE X). With reference to (TABLE III) which is an observation table, the processed data using the promethee method will assign aggregate preference functions first, note (TABLE XI). The end result of the promethee, determined from the Leaving Flow and the Entering Flow that is still partial, so combined with the mathematical process to rank, take note (TABLE XII). Comparison Results Promotion based performance performance using three methods of VIKOR, ELECTRE, and Promethee can be seen in (TABLE XIII). The results obtained provide an interpretation that the promethee method is closer to the actual results, while the vikor method is almost close to the results of truth and far away when compared with the electre method. Thus, it can be concluded that the best method for placing employee positions is the promethee method.

IV. CONCLUSION
The conclusions that can be drawn from Job Promotion for performance-based employees appear to be varied, with the VIKOR method in sequential first rank K5-K2-K3-K1-K4. with the weight in sequence is 0.0; 0.636; 0.678; 0.77, and 0.93. While the results of ranking with the method of ELECTREE in sequence is K5-K3-K2-K1-K4, with consecutive weights 2.6; 0.19; 0.14; -1.25; and -1.68. And the result of ranking with Promethee method in sequence is K5-K2-K3-K4-K1, with consecutive weights 0.371; 0.059; -0.066; -0.165, and -0.2. The results obtained provide an interpretation that the promethee method is closer to the actual results, while the vikor method is almost close to the results of truth and far away when compared with the electre method. Thus, it can be concluded that the best method for placing employee positions is the promethee method.