<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>Hryhorii Hnatiienko, Vitaliy Snytyuk, Nataliia Tmienova, Oleksii Voloshyn</article-title>
      </title-group>
      <contrib-group>
        <aff id="aff0">
          <label>0</label>
          <institution>Taras Shevchenko National University of Kyiv</institution>
          ,
          <addr-line>Volodymyrs'ka str. 64/13, Kyiv, 01601</addr-line>
          ,
          <country country="UA">Ukraine</country>
        </aff>
      </contrib-group>
      <fpage>164</fpage>
      <lpage>176</lpage>
      <abstract>
        <p>The problem of determining the effectiveness of scientific research carried out by teachers of universities is considered. In particular, the results of the publishing activity of teachers are investigated. For the comparative analysis of teachers' effectiveness, it is offered to formalize this problem in a class of problems of multicriteria optimization and to apply methods of a multiattribute choice of variants. Approaches to aggregating the effectiveness of teachers' team and justifying the advantage of choosing the most productive teams over less productive ones are also proposed. The problems of determining the most effective teacher and the most effective team of teachers are given.</p>
      </abstract>
      <kwd-group>
        <kwd>1 Scientific activity of universities</kwd>
        <kwd>teacher effectiveness</kwd>
        <kwd>multicriteria optimization</kwd>
        <kwd>weight coefficients</kwd>
        <kwd>teacher ranking</kwd>
        <kwd>aggregation of teachers' team effectiveness</kwd>
        <kwd>the most effective teacher</kwd>
      </kwd-group>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>1. Introduction</title>
      <p>- scientific research indirectly stimulates teachers to improve the quality of teaching and modernize
the disciplines they teach;</p>
      <p>- the creation of new high-quality textbooks and manuals, as a rule, is the result of the introduction
and understanding of research in universities;</p>
      <p>- the need to support at the appropriate level the scientific component of teachers and the tendency
to increase the scientific potential of university teachers, the functioning of the system of graduate and
doctoral studies is the main factors that dictate the need for scientific activities of university teachers;
- ensuring high quality of training of students, providing them with the basics of research activities
require the teaching staff to support the scientific component of their activities at a high modern level.
Of the indirect positive factors of scientific activity of teachers, which indirectly affect the teachers'
team, it should be noted that:
- scientific research promotes additional interaction between teachers and university departments;
- the need for scientific activity naturally requires the creation of new temporary research teams;
- research teams in the universities can become the basis for innovative development of new
technologies;</p>
      <p>- the exchange of ideas in the mode of scientific interaction causes the appearance of synergetic
effects and contributes to the improvement of both the scientific component of teaching and
methodological aspects of this activity.</p>
    </sec>
    <sec id="sec-2">
      <title>2. The purpose of the work</title>
      <p>The purpose of the work is to develop a mathematical model of accounting for scientific results of
teachers that ensures the relevance, completeness, and adequacy of the effectiveness of scientific
activities of teachers and teachers' teams (departments, faculties and so on).</p>
      <p>In addition, the purpose of this work is to develop a process of evaluation, accounting, and
incentives for teachers, determining the effectiveness and coordination of research, which helps to
increase the efficiency of scientific activities in domestic universities. The essential aspects of the
study of the effectiveness of scientific activity are:</p>
      <p>- control function, in particular, under the terms of the contracts concluded by the universities with
teachers;</p>
      <p>- motivation of teachers to improve the quality of methodological and research components of their
activities;</p>
      <p>- raising the rating of the universities in national and international rankings by increasing
publication activity and timely updating of information on formal achievements of teachers;
- transparency of scientific activity of teachers, opportunity to demonstrate the scientific level to
colleagues and to compare the effectiveness with other teachers;
- improvement of the corporate culture at departments and faculties.</p>
      <p>
        To ensure the quality of this area should apply a qualimetric approach [
        <xref ref-type="bibr" rid="ref5 ref6 ref7">5-7</xref>
        ] to the development of
technology for evaluating the effectiveness of the scientific work of teachers of universities. The
elements of the technology of teacher evaluation should be:
- structuring information about the state and level of the scientific subsystem of universities;
- justification of effectiveness criteria of scientific activity of teachers of the universities;
- justification and implementation of approaches to the calculation of integrated characteristics of
the scientific activity of teachers;
- development and justification of aggregate performance indicators of teachers' teams;
- introduction of modern reasonable approaches to adequate comparison of achievements of
teachers and departments.
      </p>
    </sec>
    <sec id="sec-3">
      <title>3. Objectives of research</title>
      <p>To achieve the goal of the work it is expected to perform a set of the following tasks:
- to carry out the theoretical analysis of a problem of creation of a system for an estimation and
quality control of teachers scientific activity;</p>
      <p>- consider and explore the role of the subjective component in the decision support system for the
quality of research;
- develop approaches to measuring qualimetric effectiveness indicators of research;
- to determine the main criteria for assessing the quality and effectiveness of scientific activities;
- develop and apply assessment methods using expert technologies;
- to offer and justify the interpretation of the integral effectiveness of scientific activity of teachers;
- to offer approaches to determining the ratings of teachers, taking into account the need for their
motivation to improve the quality of teaching and stimulate research.</p>
    </sec>
    <sec id="sec-4">
      <title>4. Review of recent research</title>
      <p>
        Some aspects of determining the effectiveness of scientific research have been studied by various
scientists: Belov O.V. [
        <xref ref-type="bibr" rid="ref4">4</xref>
        ], Bilukha M.T., Hnatiuk N.O., Kushnarenko N.M. [
        <xref ref-type="bibr" rid="ref1">1</xref>
        ], Surmin Yu.P. [
        <xref ref-type="bibr" rid="ref9">9</xref>
        ],
Sheiko V.M. [
        <xref ref-type="bibr" rid="ref1">1</xref>
        ] and others.
      </p>
      <p>The specificity of the study of scientific activity of teachers is that the activities of high school
teachers are multifaceted. Scientific activity, in principle, is not subject to adequate regulation, and in
cases where research is carried out on the background of teaching, the task is significantly complicated.
After all, publishing activity is only one of the many aspects of teachers' work. Therefore, the
requirements for the effectiveness of scientific activities of teachers have many features and nuances.
The complexity of studying this issue is supplemented by the fact that the scientific activity of high
school teachers is only one of the areas of their work, and not always the highest priority.</p>
      <p>At the same time, it should be noted that in the most common international and national rankings,
the most important coefficients are assigned for high research indexes:
– Times Higher Education World University Ranking;
– QS World University Ranking;
– Academic Ranking of World Universities.</p>
      <p>The national system of rating evaluation of the activity of universities is also formed taking into
account scientific indicators. Such indicators are, in particular:
- quality index of research, scientific and technical activity: "Rating of universities of Ukraine";
- indicator of international scientometric and web-metric data: "Top-200 Ukraine";
- reputation of the universities in the field of scientific research;
- the volume of the research budget of the universities per each teacher;
- teacher citation indices;
- formal indicators of scientific activity of teachers, which are reflected in the international
scientometric databases Scopus, Web of Science, etc .;
- the number of publications in the most prestigious scientific journals;
- share of foreign scientists and joint publications with foreign scientists.</p>
      <p>It should be noted that in the study of information related to research and in its analysis at the
present stage, the methods of mathematical statistics are most often used.</p>
      <p>According to the authors, to increase the level of research, appropriate mathematical models
should be developed and methods of decision theory, methods of solving multicriteria optimization
problems, expert technologies, "soft" calculations, etc. should be used in this field. This paper is
devoted to the development of this direction.</p>
    </sec>
    <sec id="sec-5">
      <title>5. Research methods</title>
      <p>In the authors’ view, to achieve the above purpose of the work and the formulated objectives, the
following scientific directions may be useful:
- multicriteria optimization problems;
- processing of expert information;
- determination of weight coefficients of criteria;
- methods of aggregation of multi-attribute data;
- methods of group choice.</p>
      <p>In determining the effectiveness of scientific activities of teachers of the universities can also be
used regulations, data from Internet resources, information about the characteristics of scientific work
of teachers, modern methods of systems analysis and information technology.</p>
      <p>
        It should be noted that there are usually different approaches to modeling decision-making
problems: multipurpose, multiattribute, and multicriteria. Multiattribute decision-making is carried
out for the problems of choosing from a set of alternatives, which are characterized by numerical
attributes, often in the presence of a single goal. Multicriteria decision-making is decision-making
with many attributes and the presence of several, usually opposite goals (criteria) [
        <xref ref-type="bibr" rid="ref10 ref11 ref12 ref13 ref14">10-14</xref>
        ].
      </p>
      <p>We will assume that the parameters of the effectiveness of teachers' research are points in a
multiattribute space. Therefore, it is logical to consider the problems of determining the effectiveness of
scientific activities of teachers as problems of multiattribute choice and formalize them in the class of
problems of multicriteria optimization.</p>
      <p>
        In addition, it is known that building a structure of preferences in a formalized form is a difficult
problem for a person: specialists in subject areas do not always have a clear idea of the structure of
preferences on the set of alternatives. In most cases, a person can not adequately determine the weight
coefficients, as well as explicitly formulate the heuristics that are used by him in the decision-making
situation [
        <xref ref-type="bibr" rid="ref15 ref16">15, 16</xref>
        ].
      </p>
      <p>
        When solving problems of multicriteria optimization, the problem of determining the area of
effective solutions is strictly objective and is solved without the use of any heuristics. Narrowing the
area of effective alternatives requires the use of additional information from experts, as effective sets
of parameters cannot be formally compared with each other. To determine a single solution to a
multicriteria problem, as a rule, three heuristics are used [
        <xref ref-type="bibr" rid="ref10">10</xref>
        ]:
      </p>
      <p>- to translate all the values of the parameters of the alternatives to the dimensionless form in a
given range of values, one of the allowable transformations is used;
- the vector of the relative importance of criteria is defined;
- it is assumed that the solution of a multicriteria problem is the point of intersection of the ray of
the normalized weight coefficients of the relative importance of the criteria and the area of effective
alternatives to the problem.</p>
      <p>We formulate heuristics that should be used to solve the problem of multicriteria optimization.</p>
      <p>Heuristics H1. The type of monotone function for converting the values of the parameters of
alternatives to the dimensionless form is carried out according to formulas that must meet the
following requirements:
- take into account the need to minimize deviations from the optimal values for each parameter;
- have a common starting point and the same order of change of values on the whole set of
alternatives;</p>
      <p>- maintain the preference ratio on the set of alternatives being compared, according to the set of
parameters, and thus not change the set of effective alternatives.</p>
      <p>Heuristics H2. The best alternative in solving the problem of multicriteria optimization should be
considered an alternative for which deviations from the best values of the parameters for each
estimate are minimal.</p>
      <p>To achieve the goal of the research formulated in this paper, we will also use expert technologies
that are widely used in various fields of human life and are actively developing in recent decades.</p>
      <p>One of the key concepts of expert technology is heuristics, which can be axioms, postulates,
assumptions, presumptions, paradigms, hypotheses, additions, propositions, and so on. Heuristics are
empirical methodological rules that can help to find solutions and help to define incorrect problems.</p>
    </sec>
    <sec id="sec-6">
      <title>6. Mathematical model of scientific effectiveness</title>
      <p>The peculiarity of the scientific activity of teachers is that such activity, as a rule, is not exclusive
or autonomous. The teacher of universities should carry out research on the background of his other
professional activities. Therefore, the research of the teacher, on the one hand, can not be separated
from the general flow of its activities, on the other hand, research, directly and indirectly, has a great
influence on the other areas of activity of the teacher.</p>
      <p>Let us build a table that characterizes and illustrates the criterion base for evaluating the
effectiveness of scientific activities of teachers of universities.</p>
      <sec id="sec-6-1">
        <title>Execution of academic Supervision in</title>
        <p>workload: lectures, student groups.
seminars, practical classes,
laboratory work, etc. Preparation of
Ensuring the educational student
process and research area: competitions.
management of diplomas, Organization and
term papers, etc. holding of
Development of curricula, competitions of
work programs of academic student scientific
disciplines. works.</p>
      </sec>
      <sec id="sec-6-2">
        <title>Organization of hackathons, etc.</title>
      </sec>
      <sec id="sec-6-3">
        <title>Certification training Interaction with and raising the leadership and personal scientific colleagues level</title>
      </sec>
      <sec id="sec-6-4">
        <title>Certification training Execution of</title>
        <p>courses. standing</p>
      </sec>
      <sec id="sec-6-5">
        <title>Acquaintance with assignments. the latest scientific achievements.</title>
      </sec>
      <sec id="sec-6-6">
        <title>Participation in the meetings of the department, etc.</title>
        <p>Let the number of teachers whose activities should be compared be equal to n , and let the set of
their indices be denoted by J  1,..., n .</p>
        <p>To prepare a mathematical model for determining the effectiveness of scientific activities of
teachers, let us consider some attributes of scientific and practical activities of teachers of universities
vi  (v1i , v2i , v3i, v4i , v5i , v6i , v7i , v8i , v9i ), i  J , which can be used to determine the effectiveness of their
research and used to compare the effectiveness of different teachers, different departments, as well as
to calculate the dynamics of this indicator in different periods:
v1i – the total number of publications of teachers, i  J ;
v2i – the number of publications of teachers in a given period, i  J ;
v3i – the total number of publications of teachers in international scientometric databases (Scopus,
Web of Science, etc.), i  J ;</p>
        <p>v4i – the number of publications of teachers in international scientometric databases (Scopus, Web
of Science, etc.) in a given period, i  J ;</p>
        <p>v5i – index of citations of scientific works of teachers (total number of citations of all scientific
works of the teacher, department, faculty or university in scientometric bases), i  J ;
v6i – index of citations of scientific works of teachers in a given period (total number of citations
for the period of all the above scientific papers), i  J ;</p>
        <p>v7i – average index of citations of scientific works of teachers (average number of citations per one
scientific work of a teacher, department, faculty or university in scientometric databases), i  J ;
v8i – average index of citations of scientific works of teachers in a given period (average number of
citations for the period per one scientific work of a teacher, department, faculty or university in
scientometric databases), i  J ;</p>
        <p>The problems of determining the most effective teacher (from the point of view of scientific
activity), the most effective department in the scientific activity, or the most productive faculty in the
scientific direction will be formalized in the class of multicriteria optimization problems. In this case,
taking into account the need to use heuristics in such cases, we will pay considerable attention to the
subjective component of multicriteria problems.</p>
        <p>Note that today there are three main approaches to describing the problems of introspective
(internal, in-depth) analysis: using binary relations, the choice function, and the criterion approach.
The latter approach involves the assumption that each alternative can be evaluated by a specific
number, which is the value of the criterion, so the comparison of alternatives is reduced to comparing
the corresponding numbers. It is clear that in many practical situations, multicriteria is a way to
increase the adequacy of goal description.</p>
        <p>The problem of multicriteria optimization is formalized in the following formulation:
fi v  max, i  I1,
fi v  min, i  I 2 ,
v  A, A  E k ,
where A  set of alternatives (in our case – indicators of scientific activity of teachers), which
are characterized by k parameters, which belong to space E k ;</p>
        <p>y(v)  ( f1(v),..., fk (v)) – vector of evaluations of alternatives or criteria, which is specified by the
mapping f : A  E k ;</p>
        <p>I  1,..., k set of indexes of criteria;</p>
        <p>I1  1,..., k1, I 2  k1  1,..., k – sets of indices of criterion functions, which, respectively, are
maximized and minimized, I1  I2  I.</p>
        <p>Thus, when solving the problem of determining the best teacher, the following approach can be
proposed. First, the integrated indicators of each teacher are determined by some aspects (for
example, popularity among students, publishing activity, social and organizational activities, and so
on). Moreover, the application of this indicator may be preceded by stratification, namely the
distribution of voting participants to bachelors, graduates, only those who attended classes more often,
and so on. At the next stage, an aggregate indicator of publishing activity is determined among
teachers or teachers' teams. Other aggregate indicators are also calculated. The winner is then
determined by solving the multicriteria optimization problem.</p>
        <p>That is, in most cases, the solution of problems related to the analysis of teachers' activities is a
compromise. Note that for the solution adequate finding and justification should provide the
calculation of the weight coefficients of the criteria.</p>
      </sec>
    </sec>
    <sec id="sec-7">
      <title>7. Determination of weight coefficients of criteria</title>
      <p>It is possible at the initial stage to determine the weight coefficients by experts, although it is
known that a person can not reliably assign weight. Therefore, it is necessary to comprehensively
approach and determine by indirect methods the importance of parameters that characterize the
effectiveness of scientific activities of teachers.</p>
      <p>
        Research on expert evaluation problems and the practice of building decision support systems
show that experts and decision-makers do not always have a clear idea of the structure of preferences
on the set of alternatives [
        <xref ref-type="bibr" rid="ref16">16</xref>
        ]. In most cases, a person can not adequately determine the weight
coefficients, as well as to allocate in the explicit case of heuristics, which are used by him in the
decision-making situation.
      </p>
      <p>There are several common ways to represent the values of the weight coefficients n of the criteria
of the problem of type (1):
(1)
- arbitrary real or natural numbers     i  , i  I ;
- real numbers, taking into account restrictions (one-sided or two-sided), for example,  i  0, i  I;
 5   i  5, i  I; 0   i  1, i  I ;
  i  0,    i  , i  I ;
iI
real
or
natural
numbers,
taking
into
account
the
condition
of
centering:
- real numbers taking into account the condition of normalization:  i  1,  i  0, i  I ;
iI
- real numbers taking into account the idealization condition: miaIx  i  1,  i  0, i  I .</p>
      <p>A common form of representation of normalized weight coefficients is the interval form
   iH ,  iB  , i  I , 0  iH  iB  1, i  I.</p>
      <p>
        The method of determining the function of belonging of the weight coefficients values to the fuzzy
set (0,1) is also used. Approaches to the definition of membership functions and algorithms for
constructing membership functions based on the analysis of the frequency of the values are that each
weight coefficient as a result of the procedure of accounting for the frequency of values will be
characterized by its membership function to fuzzy set [
        <xref ref-type="bibr" rid="ref17">17</xref>
        ]. This approach significantly expands the
possibilities of modeling subject areas and solving problems of multicriteria optimization [
        <xref ref-type="bibr" rid="ref18">18</xref>
        ].
      </p>
    </sec>
    <sec id="sec-8">
      <title>8. Aggregation of effectiveness of scientific activity of teachers</title>
      <p>Decision-making is based on individual or group introspective analysis of the problem and the choice
of the way how to solve it. The introspective analysis consists in the observation of the researcher's own
feelings, thoughts, images, experiences, acts of thinking without the use of any tools or standards and is
not accompanied by significant loss of information. The study of introspective analysis and processing
of the subjective component in the application of expert technologies in poorly structured complex
systems is an important and relevant direction for improving decision support processes.</p>
      <p>The problems of analysis of the scientific activity of teachers can be adequately formalized in the
class of problems of ranking alternatives that is ordering the set of alternatives according to the degree of
manifestation of some properties. An important and widespread tool for the application of expert
technologies, which is now a classic, is the task of collective determination of the ranking of
alternatives, which by some criteria is "closest" to all rankings built by analyzing parameters. The most
reasonable method of finding the resulting ranking of alternatives is to calculate the median of the given
rankings.</p>
      <p>It should be noted that the aggregation of the effectiveness of scientific activities of teachers can be
carried out using different measurement scales: absolute, ratio, interval, ordinal, or nominal. One of the
important heuristics in the problems of introspective analysis of the subjective component of
decisionmaking is to determine the average that should be applied. This is due to the fact that an important place
among all methods of data analysis is occupied by data averaging algorithms. Today, there are several
common averages used to analyze subjective information.</p>
      <p>
        It is known [
        <xref ref-type="bibr" rid="ref19">19</xref>
        ] that for nominal features, namely measured in the scale of names, the only mode is an
acceptable average. For data measured on an ordinal scale, the median is acceptable. When examining the
information measured in the interval scale, it is possible to use only the arithmetic mean. And for the
analysis of the data set in the scale of relations, degree averages and geometric averages are used.
      </p>
    </sec>
    <sec id="sec-9">
      <title>9. Metrics for measuring the distances between the indicators of scientific activity of teachers</title>
      <p>When comparing the effectiveness of scientific activities of teachers of universities, each teacher
corresponds to a point in the parametric space
vi  v1i ,..., vki  V , i  J  1,..., n , I  1,..., k,
(2)
where n  the number of teachers, k  the number of assessment parameters.</p>
      <p>When organizing the achievements of teachers, the absolute values of their scientific achievements
can be used or the achievements of teachers can be compared on an ordinal (rank) scale. That is,
teachers can compare with each other on different values of parameters in the cardinal and ordinal
scales. If the parameters of type (2) cannot be measured objectively, then expert estimates are used.
Ordering teachers’ achievements according to the values of parameters that cannot be measured in
quantitative scales can be formalized in the class of collective ranking problems.</p>
    </sec>
    <sec id="sec-10">
      <title>9.1. Formalization of the problem of ranking teachers' achievements in assessment in ordinary measurement scales</title>
      <p>In practical situations, problems when the parameters of teachers cannot or should not be measured
in absolute terms often arise. Therefore, it is advisable to consider only the relationship between
performance indicators of teachers that is to set the ranking of teachers' achievements on those
parameters that are considered important for comparison or important for some current research.</p>
      <p>Since teachers are compared by a group of parameters, the problem to determine the resulting
ranking of teachers arises (taking into account the whole set of parameters). The most common
method of finding a compromise ranking in group selection problems is to calculate the median of the
given rankings. This group of methods used to summarize expert information is the most reliable and
mathematically sound.</p>
    </sec>
    <sec id="sec-11">
      <title>9.1.1. Statement of the problem of determining the resulting ordering of teachers by a group of parameters specified in the ordinal scales</title>
      <p>Suppose that k parameters are given, for each of which it is possible to evaluate teachers, that is to
build such a ranking of teachers for each parameter, which indicates the degree of manifestation of
this parameter in the teacher's activity. The smaller the value of the selected parameter has the teacher,
the lower will be his rank in a given ranking</p>
      <p>Ri  ai1 ≽...≽ ain , i j  J  1,..., n, j  J , i  I. (3)
It is necessary to find some group (resultant, aggregate, collective, consensus, integrative) ranking of
n teachers R*  ai1 ≽...≽ ain , i j  J , j  J , which will be the closest in some sense to the
rankings of teachers of type (3), built taking into account each parameter.</p>
      <p>
        In this paper, the symbol ≽ denotes the relation of non-strict advantage, namely, when
≽  ,  . Thus, the problem is logically formalized in the class of problems of non-strict
collective ranking [
        <xref ref-type="bibr" rid="ref20">20</xref>
        ] (definition of perfect quasi-orders [
        <xref ref-type="bibr" rid="ref21">21</xref>
        ], ordering [
        <xref ref-type="bibr" rid="ref22">22</xref>
        ], quasi-series [
        <xref ref-type="bibr" rid="ref23">23</xref>
        ],
ranking with connections [
        <xref ref-type="bibr" rid="ref24">24</xref>
        ], quasi-orders, clustered rankings [
        <xref ref-type="bibr" rid="ref25">25</xref>
        ]).
      </p>
    </sec>
    <sec id="sec-12">
      <title>9.1.2. Statement of the problem of determining the resulting ordering of teachers by a group of parameters specified in the ordinal scales with incomplete rankings</title>
      <p>
        In practice, there are often situations when not every teacher can be evaluated on all the parameters
selected for evaluation. Requiring a mandatory assessment of all teachers means deliberately creating
inaccurate initial data. Therefore, it is necessary to forecast situations when for each parameter it is
possible to establish a partial order on the subset of teachers Ai , i  I , the whole set of teachers
A, Ai  A, i  I , for whom this is possible. It should be noted that the problems associated with
incomplete data have been studied by many scientists, in particular, in [
        <xref ref-type="bibr" rid="ref26 ref27">26, 27</xref>
        ].
      </p>
      <p>We introduce the definition of the incomplete ranking of teachers RiH , i  I , of the set A : it is a
binary relation given on a subset of teachers A, A  A , which satisfies the properties of
completeness, reflexivity, antisymmetry, transitivity: but only on the subset A, A  A , and not on
the whole set A .</p>
      <p>Let be given partial orders by k parameters</p>
      <p>RHi  ai1 ≽...≽ ain , i j  J  1,..., n, j  J , ni  n, i  I , (4)
on the selected subsets of teachers Ai  A,i  I. It is necessary to find some resulting ordering of n
teachers R*  ai1 ≽...≽ ain , i j  J , j  J , which is built on the basis of given incomplete orders of
the form (4).</p>
      <p>The initial stage of solving the problem described in this paper is to unite all the teachers, whose
achievements are arranged according to different parameters, into a single set ai  A, i  1,..., n,
to determine the resulting ranking R* . That is, by the subsets Ai  A, i  I , the full set of teachers
k
i1
A </p>
      <p>Ai is determined.</p>
      <p>It is clear that different variants of relations between subsets are allowed: Ai1  Ai2
 ,
Ai1  Ai2  , Ai1  Ai2 , i1,i2  I.</p>
      <p>The set of all possible teacher rankings by parameters is an area of acceptable solutions for
determining the resulting ranking of teachers R* , built on incomplete rankings of the form (4).</p>
    </sec>
    <sec id="sec-13">
      <title>9.2. Metrics for measuring the scientific activity of teachers distances between the indicators of</title>
      <p>When applying the algebraic approach, metrics are introduced to measure distances.
1) Cook metric of mismatch of ranks (places, positions) of teachers in rankings by each of the
parameters
d R j , Rl    ri j  ril ,</p>
      <p>iI
l
where ri - the rank of the i -th teacher in the ranking by the l -th parameter of assessment
Rl , l  L, 1  ril  n.</p>
      <p>Note that the values of the ranks of teachers may not be integers, as the problem is formalized in
the class of group finding a non-strict ranking.</p>
      <p>2) Hamming metric presupposes a transition to another space. To move from the space of ranks to
the space of pairwise comparisons of teachers, individual preferences for each of the parameters are
presented in the form of a matrix of pairwise comparisons</p>
      <p>Bl  bilj , j  I , l  L, (6)
where bilj  1, i, j  I , l  L , if and only if the і-th teacher dominates the j-th teacher by the l-th
parameter. Moreover, bilj  blji , i, j  I ,l  L. If the values of the parameters specified in the ordinal
scale are equivalent for two teachers, then bilj  blji  0, i, j  I ,l  L .</p>
      <p>Hamming metric is used to determine the distances between teachers’ relationships
d H (B j , Bl )  0,5  bisj  bils , j, l  I , i, s  J .</p>
      <p>iI sI
3) The quadratic metric looks like this
d k (B j , Bl )     bisj  bils 2 

 iI sI 
4) The metric of dominance are also used:
1/ 2</p>
      <p>, j, l  I , i, s  J .
d d (B j , Bl )  max bisj  bils , j, l  I , i, s  J .</p>
      <p>
        i,sI
(5)
l  I , where  il  the number of indicators of teachers, which precede the indicators of the i  th
teacher in the l  th ranking of the form (4). B.G. Litvak [
        <xref ref-type="bibr" rid="ref28">28</xref>
        ] proposed for the vectors of preferences
 1 and  2 , formed on the basis of rankings R1 and R2 , to determine the distance by the formula:
d  (R1, R 2 )    i1  i2 .
      </p>
      <p>iI</p>
    </sec>
    <sec id="sec-14">
      <title>9.3. Criteria for determining the generalized ranking of achievements on the basis of rankings set by individual parameters teachers'</title>
      <p>For the Cook metric (5) using the utilitarian criterion, the Cook-Sayford median is calculated:
RCS  CS  Arg min  d K R, Rl .</p>
      <p>RR lL
When using the egalitarian criterion GV -median (compromise) is calculated:</p>
      <p>RGV  GV  Arg min max d K R, Rl .</p>
      <p>RR lL
For the Hamming metric (6) using the utilitarian criterion, the Kemeni-Snell median is calculated:
R КС  КС  Arg min  d H B, Bl .</p>
      <p>BB lL
When using the egalitarian criterion VG -median (compromise) is calculated:</p>
      <p>RVG  VG  Arg min max d H B, Bl .</p>
      <p>BB lL</p>
      <p>Similar criteria that determine the medians of given rankings are used for quadratic metric,
dominance metric, and preference vector.</p>
    </sec>
    <sec id="sec-15">
      <title>9.4. Peculiarities of taking into account incomplete rankings of teachers, determined by parameters in ordinal scales, when calculating the generalized ranking of teachers’ achievements</title>
      <p>Taking into account the peculiarities of incomplete rankings of teachers by the parameters that
characterize their activities, requires the introduction of additional heuristics.</p>
      <p>Heuristics Н3. The distance from any ranking R* to each ranking, determined by comparing the
results of scientific activities of teachers, RiН ,i  I , is equal to the sum of the probabilistic and definite
part.</p>
      <p>Heuristics Н4. The absence of assessments of individual teachers by some parameters creates an
unknown relations between all other assessments of teachers and does not participate in the ranking
that is this assessment of the teacher is not represented in the incomplete ranking. Thus, when
specifying incomplete rankings for assessments for each parameter, we have the following number of
teachers’ assessments:</p>
      <p>ni  given by the parameter i  I in the incomplete ranking RiН ,i  I , which will be a definite
part of the distances;</p>
      <p>(n  ni )  i  not specified by the parameter
RiН ,i  1,..., k, which will be a probabilistic part of the distances.
i  I
in
the
incomplete
ranking</p>
      <p>Individual incomplete advantages given by each parameter on subsets of teachers' assessments
RlН , l  І , can also be represented as an incomplete matrix of pairwise comparisons (MPC)</p>
      <p>BlН  biljН , j  I , l  L.</p>
      <p>Hamming metric is used to determine the distances between the relations of this matrix.</p>
      <p>Heuristics E5. The mathematical expectation of indefinite distances between teachers' assessments
in the ranking is equal to 8/9. That is, the distance between the elements of the MPC, at least one of
which is not defined, must be equal to 8/9 (based on the assumption that the equality of its values
"-1", "0" or "1" are equally likely). And the distances between the elements have respectively the
following distributions: (0,1,2), (1,0,1), (2,1,0). The probabilistic part from ranking R iН , i  I , given
by i  th ( i  I ) parameter to any other ranking for Hamming
metric is always equal
9  i  ( i 1) / 16, i  I .</p>
      <p>10. Aggregation of data in the calculation of objective indicators of
scientific activity of teachers, defined in the cardinal scales</p>
      <p>The accumulated experience of expert evaluation in various areas of human activity convincingly
shows that any statistical operations become more useful and reasonable when reducing the number of
features used for analysis. Therefore, the problem of aggregating the features that characterize the activities
of teachers to a smaller number of constructed "factors" (aspects, etc.) occupies a significant place in the
problems of determining the effectiveness of scientific activities of teachers. The analysis of the set of
teachers' assessments by a group of parameters is to determine the level of general consistency of teachers'
assessments and to select, if necessary, a group of "homogeneous" subgroups that combine the parameters
of teachers with agreed assessments. The formulation of these problems is dictated by the fact that the
transition to the aggregation of estimates by different parameters is possible only after identifying the
structure of preferences. For example, if the overall consistency of estimates by parameters is low and the
group of parameters is divided into several subgroups, within which the consistency of estimates is high,
then aggregation should be performed for these subgroups by estimates of parameters.</p>
      <p>In the analysis of assessments of scientific activity of teachers and in determining the relative
importance of publications, there are problems of presenting these assessments in a systematic way and
there are problems of comparison and aggregation of assessments. The use of mathematical methods in
the analysis of expert assessments allows to adequately summarize the judgments of specialists and
identify the information that they have in a latent form.</p>
      <p>If there are objective parameters that characterize the activities of teachers, different approaches can
be proposed. We introduce the notation for the parameters of scientific activity of each teacher
j, j  J , as follows:</p>
      <p>v1j  the number of articles in editions that are included in the scientometric databases of Scopus or
WoS by the j  th teacher;</p>
      <p>v2j  the number of articles in conference proceedings, which are included in the scientometric
databases of Scopus or WoS by the j  th teacher;
v3j  the number of articles in Ukrainian periodicals by the j  th teacher;
v4j  the number of abstracts at international and Ukrainian conferences by the j  th teacher;
v5j  the number of training manual and textbooks by the j  th teacher;
v6j  the number of monographs by the j  th teacher;
v7j  the number of prize-winning students of All-Ukrainian Olympiads and All-Ukrainian
competitions of scientific works (2nd round), the supervisor of which is a j  th teacher;
k s  the number of co-authors in the relevant scientific work (the author is one of the co-authors).</p>
      <p>To determine the integrated (aggregated) effectiveness of the scientific activity of the j  th teacher
on the basis of objective indicators, it is proposed to use the following empirically derived formula:
 v1j v2j v3j  v4j  v5j v6j 
Q1   6 1 / ks   3 1 / ks    1 / ks    s1 1 / ks   / 5  41 / ks  81 / ks  5v7j  /
j  s1 s1 s1  s1 s1 
 v1i v2i v3i  v4i  v5i v6i 
max  61/ ks   31/ ks   1/ ks    s1 1/ ks  / 5  41/ ks  81/ ks  5v7i  , j  J .
iJ  s1 s1 s1  s1 s1 
We will also introduce indicators related to the survey of students to determine the best teacher:
v8j  the number of students who voted for the j  th teacher;
v9j  the number of students who took part in the voting;
v1j0  the number of students studying at the department where the j  th teacher works.
Q2  v 8j /v9j  v8j / v1j0  / max v 8i /v9i  vi / v1i0 </p>
      <p>j iI 8
The integral value of the teacher's rating will be determined by the formula:</p>
      <p>Q3  Qj1 / 2  Q2 / 2, j  J .</p>
      <p>j j
12. Conclusions</p>
      <p>In this work, the approaches to research of productivity of scientific activity of teachers of
universities, comparison of scientific activity of departments, faculties, and universities as a whole are
investigated. The result of scientific research is a publication that confirms the fact of scientific
accomplishment, with which the scientist will be able to familiarize not only his colleagues but also
the world community. The scientific work is not completed until it is published and indexed in a
scientometric database. In the modern scientific world, publishing activity is becoming increasingly
important for every scientist and teacher of universities, regardless of the direction of his research.
The following main scientific results are obtained:
- the statement of problems of definition of scientific researches productivity of teachers is offered;
- a theoretical analysis of the problem of creating a system of evaluation and quality control of
scientific activities of teachers is carried through;</p>
      <p>- the role of the subjective component in the decision support system regarding the quality of
scientific research is considered and researched;
- approaches to measuring qualimetric performance indicators of scientific research is developed;
- criteria for assessing the quality and effectiveness of scientific activities are defined;
- approaches to determining the ratings of teachers are proposed, taking into account the need for
their motivation;</p>
      <p>- the interpretation of the integral efficiency of scientific activity of teachers is offered and
substantiated.</p>
      <p>In the future, based on the analysis of the obtained solutions to the problem of determining the
results of scientific research of teachers, approaches to determining the coefficients of the relative
performance of teachers in the form of membership function to a fuzzy set can be proposed.</p>
      <p>13. References</p>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          [1]
          <string-name>
            <surname>Sheiko</surname>
            ,
            <given-names>V. M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Kushnarenko</surname>
            ,
            <given-names>N. M.</given-names>
          </string-name>
          <article-title>Organization and methods of research activities: a textbook for higher education / V.M. Sheiko</article-title>
          ,
          <string-name>
            <given-names>N.M.</given-names>
            <surname>Kushnarenko - X: KhDAK</surname>
          </string-name>
          ,
          <year>1998</year>
          . - 288 p. (
          <year>1998</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          <source>[2] Methodology and organization of scientific research: Lecture notes / Compiler VM Sour</source>
          . - Sumy: Sumy State University Publishing House,
          <year>2009</year>
          .-113 p. (
          <year>2009</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref3">
        <mixed-citation>
          <article-title>[3] National report on the state and prospects of education in Ukraine / Nat. acad. ped</article-title>
          . Sciences of Ukraine; [editor: V.G.
          <article-title>Kremen (chairman), V.I. Lugovyi (deputy chairman), A.M. Gurzhiy (deputy chairman), O. Ya. Savchenko (deputy chairman)]; for general ed</article-title>
          . V.G. Kremenya. - Kyiv: Pedagogical Thought,
          <year>2016</year>
          . - 448 p. (
          <year>2016</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref4">
        <mixed-citation>[4] State Statistics Service of Ukraine: http://www.ukrstat.gov.ua/</mixed-citation>
      </ref>
      <ref id="ref5">
        <mixed-citation>
          [5]
          <string-name>
            <surname>Azgaldov</surname>
            ,
            <given-names>G.G.</given-names>
          </string-name>
          and
          <string-name>
            <surname>Kostin</surname>
            ,
            <given-names>A.V.</given-names>
          </string-name>
          <article-title>Intellectual property, innovation</article-title>
          and qualimetry // Economic strategies,
          <year>2008</year>
          . - Vol.
          <volume>2</volume>
          . - pp.
          <fpage>162</fpage>
          -
          <lpage>164</lpage>
          . (
          <year>2008</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref6">
        <mixed-citation>
          [6]
          <string-name>
            <surname>Azgaldov</surname>
            ,
            <given-names>G.G.</given-names>
          </string-name>
          and
          <string-name>
            <surname>Kostin</surname>
            ,
            <given-names>A.V.</given-names>
          </string-name>
          (
          <year>2011</year>
          ),
          <article-title>"Applied Qualimetry: Its origins, errors and misconceptions"</article-title>
          ,
          <source>Benchmarking: An International Journal</source>
          , Vol.
          <volume>18</volume>
          Iss: 3, pp.
          <fpage>428</fpage>
          -
          <lpage>444</lpage>
          . (
          <year>2011</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref7">
        <mixed-citation>
          [7]
          <string-name>
            <surname>Per</surname>
            ,
            <given-names>Yu.S.</given-names>
          </string-name>
          <article-title>Economic metrology. Qualimetry in design management: Monograph. - Izhevsk: Publishing Center "Udmurt University"</article-title>
          .
          <year>2019</year>
          . - 448 p. (
          <year>2019</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref8">
        <mixed-citation>
          [8]
          <string-name>
            <surname>Belov</surname>
            ,
            <given-names>O. V.</given-names>
          </string-name>
          <article-title>Comparative analysis of science funding in the world: trends and ratings</article-title>
          . - Scientific Bulletin of Poltava University of Economics and Trade. -
          <year>2014</year>
          . - Vol.
          <volume>5</volume>
          (
          <issue>67</issue>
          ). (
          <year>2014</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref9">
        <mixed-citation>
          [9]
          <string-name>
            <surname>Surmin</surname>
            ,
            <given-names>Yu. P.</given-names>
          </string-name>
          <article-title>Modernization of scientific knowledge management in Ukraine as a prerequisite for building a knowledge society / Yu</article-title>
          . P. Surmin. // Bulletin of the
          <source>National Academy of Public Administration under the President of Ukraine</source>
          . - Vol.
          <volume>3</volume>
          . -
          <fpage>2013</fpage>
          . - P.
          <fpage>98</fpage>
          -
          <lpage>105</lpage>
          . (
          <year>2013</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref10">
        <mixed-citation>
          [10]
          <string-name>
            <surname>Mikhalevich</surname>
            ,
            <given-names>V. S.</given-names>
          </string-name>
          and
          <string-name>
            <surname>Volkovich</surname>
          </string-name>
          , V. L.
          <article-title>Computational methods of research and design of complex systems</article-title>
          . - M .:
          <string-name>
            <surname>Nauka</surname>
          </string-name>
          ,
          <year>1982</year>
          . - 286 p. (
          <year>1982</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref11">
        <mixed-citation>
          [11]
          <string-name>
            <surname>Volkovich</surname>
            ,
            <given-names>V. L.</given-names>
          </string-name>
          <article-title>Problems of creating intelligent decision support systems</article-title>
          . - K.,
          <year>1990</year>
          . - 190 p. (
          <year>1990</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref12">
        <mixed-citation>
          [12]
          <string-name>
            <surname>Zgurovsky</surname>
            ,
            <given-names>M.Z.</given-names>
          </string-name>
          <article-title>Models and methods of decision making in fuzzy conditions / M. Z</article-title>
          .
          <string-name>
            <surname>Zgurovsky</surname>
          </string-name>
          .
          <article-title>-</article-title>
          K.:
          <string-name>
            <surname>Nauk</surname>
          </string-name>
          . Dumka,
          <year>2011</year>
          . - 352 p. (
          <year>2011</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref13">
        <mixed-citation>
          [13]
          <string-name>
            <surname>Zaichenko</surname>
            ,
            <given-names>E. Yu.</given-names>
          </string-name>
          and
          <string-name>
            <surname>Zaichenko</surname>
            , Yu.,
            <given-names>P.</given-names>
          </string-name>
          <article-title>Multi-criteria decision-making problems in fuzzy conditions / E</article-title>
          . Yu. Zaichenko, Yu. P. Zaichenko // Systemic information and information technologies: international science and
          <source>technology journal. - 2016</source>
          . - Vol.
          <volume>4</volume>
          . - P.
          <fpage>79</fpage>
          -
          <lpage>87</lpage>
          . (
          <year>2016</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref14">
        <mixed-citation>
          [14]
          <string-name>
            <surname>Tsyganok</surname>
            ,
            <given-names>V.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Kadenko</surname>
            ,
            <given-names>S.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Andriychuk</surname>
            ,
            <given-names>O.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Roik</surname>
            ,
            <given-names>P.</given-names>
          </string-name>
          <article-title>Usage of multicriteria decision-making support arsenal for strategic planning in environmental protection sphere</article-title>
          .
          <source>Journal of MultiCriteria Decision Analysis</source>
          <volume>24</volume>
          (
          <issue>5-6</issue>
          ),
          <year>2017</year>
          , pp.
          <fpage>227</fpage>
          -
          <lpage>238</lpage>
          .
        </mixed-citation>
      </ref>
      <ref id="ref15">
        <mixed-citation>
          [15]
          <string-name>
            <surname>Hnatiienko</surname>
            <given-names>H. Choice</given-names>
          </string-name>
          <article-title>Manipulation in Multicriteria Optimization Problems / Selected Papers of the XIX International Scientific and Practical Conference "Information Technologies and Security"</article-title>
          (ITS
          <year>2019</year>
          ), pp.
          <fpage>234</fpage>
          -
          <lpage>245</lpage>
          (
          <year>2019</year>
          ).
        </mixed-citation>
      </ref>
      <ref id="ref16">
        <mixed-citation>
          [16]
          <string-name>
            <surname>Tsyganok</surname>
            <given-names>V.V.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Kadenko</surname>
            <given-names>S.V.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Andriichuk</surname>
            <given-names>O.V.</given-names>
          </string-name>
          <string-name>
            <surname>Considering</surname>
          </string-name>
          <article-title>Importance of Information Sources during Aggregation of Alternative Rankings /</article-title>
          <source>CEUR Workshop Proceedings</source>
          , Vol.
          <year>2067</year>
          .
          <article-title>Selected Papers of the XVII International Scientific</article-title>
          and Practical Conference on Information Technologies and
          <string-name>
            <surname>Security (ITS</surname>
          </string-name>
          <year>2017</year>
          ); Kyiv, Ukraine, November
          <volume>30</volume>
          ,
          <year>2017</year>
          . P.
          <volume>132</volume>
          -
          <fpage>141</fpage>
          .
        </mixed-citation>
      </ref>
      <ref id="ref17">
        <mixed-citation>
          [17]
          <string-name>
            <surname>Hnatiienko</surname>
            <given-names>H.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Snytyuk</surname>
            <given-names>V.</given-names>
          </string-name>
          <article-title>A posteriori determination of expert competence under uncertainty / Selected Papers of the XIX International Scientific and Practical Conference "Information Technologies and Security"</article-title>
          (ITS
          <year>2019</year>
          ), pp.
          <fpage>82</fpage>
          -
          <lpage>99</lpage>
          (
          <year>2019</year>
          ).
        </mixed-citation>
      </ref>
      <ref id="ref18">
        <mixed-citation>
          [18]
          <string-name>
            <surname>Samokhvalov</surname>
            <given-names>Y.</given-names>
          </string-name>
          (
          <year>2020</year>
          )
          <article-title>Construction of the Job Duration Distribution in Network Models for a Set of Fuzzy Expert Estimates</article-title>
          . In: Lytvynenko V.,
          <string-name>
            <surname>Babichev</surname>
            <given-names>S.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Wójcik</surname>
            <given-names>W.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Vynokurova</surname>
            <given-names>O.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Vyshemyrskaya</surname>
            <given-names>S.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Radetskaya</surname>
            <given-names>S</given-names>
          </string-name>
          .
          <source>(eds) Lecture Notes in Computational Intelligence and Decision Making. ISDMCI 2019. Advances in Intelligent Systems and Computing</source>
          , vol
          <volume>1020</volume>
          . Springer, Cham. https://doi.org/10.1007/978-3-
          <fpage>030</fpage>
          -26474-
          <issue>1</issue>
          _
          <fpage>8</fpage>
        </mixed-citation>
      </ref>
      <ref id="ref19">
        <mixed-citation>
          [19]
          <string-name>
            <surname>Orlov</surname>
            ,
            <given-names>A.I.</given-names>
          </string-name>
          <article-title>Decision making theory</article-title>
          . Textbook /
          <string-name>
            <given-names>A. I.</given-names>
            <surname>Orlov</surname>
          </string-name>
          . - M.: Exam Publishing House,
          <year>2005</year>
          . - 656 p. (
          <year>2005</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref20">
        <mixed-citation>
          [20]
          <string-name>
            <surname>Makarov</surname>
            ,
            <given-names>I.M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Vinogradskaya</surname>
            ,
            <given-names>T.M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Rubchinsky</surname>
            ,
            <given-names>A.A.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Sokolov</surname>
          </string-name>
          , V.V.
          <article-title>Theory of choice and decision making</article-title>
          . Moscow: Nauka,
          <year>1982</year>
          . - 330 p. (
          <year>1982</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref21">
        <mixed-citation>
          [21]
          <string-name>
            <surname>Shreider</surname>
            ,
            <given-names>Yu.A.</given-names>
          </string-name>
          <string-name>
            <surname>Equality</surname>
          </string-name>
          , similarity, order. - M.:
          <string-name>
            <surname>Nauka</surname>
          </string-name>
          ,
          <year>1971</year>
          . - 256 p. (
          <year>1971</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref22">
        <mixed-citation>
          [22]
          <string-name>
            <surname>Kemeny</surname>
            ,
            <given-names>J.</given-names>
          </string-name>
          and
          <string-name>
            <surname>Snell</surname>
            ,
            <given-names>J. Cybernetic Modeling: Some</given-names>
          </string-name>
          <string-name>
            <surname>Applications</surname>
          </string-name>
          . - M .: Soviet radio,
          <year>1972</year>
          . - 192 p. (
          <year>1972</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref23">
        <mixed-citation>
          [23]
          <string-name>
            <surname>Mirkin</surname>
            ,
            <given-names>B.G.</given-names>
          </string-name>
          <article-title>Group choice problem</article-title>
          . Moscow: Nauka,
          <year>1973</year>
          . - 256 p. (
          <year>1973</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref24">
        <mixed-citation>
          [24]
          <string-name>
            <surname>Hollender</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          and
          <string-name>
            <surname>Wolfe</surname>
          </string-name>
          , D. Nonparametric methods of statistics. Moscow: Finance and Statistics,
          <year>1983</year>
          . - 518 p. (
          <year>1983</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref25">
        <mixed-citation>
          [25]
          <string-name>
            <surname>Orlov</surname>
            <given-names>A.I.</given-names>
          </string-name>
          <article-title>Methods for making managerial decisions: textbook - M .:</article-title>
          <string-name>
            <surname>KNORUS</surname>
          </string-name>
          ,
          <year>2018</year>
          . - 286 p. (
          <year>2018</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref26">
        <mixed-citation>
          [26]
          <string-name>
            <surname>Tsyganok</surname>
            <given-names>Vitalii</given-names>
          </string-name>
          , Kadenko Sergii, Andriichuk Oleh,
          <article-title>Roik Pavlo Combinatorial Method for Aggregation of Incomplete Group Judgments</article-title>
          .
          <source>Proceedings of 2018 IEEE First International Conference on System Analysis &amp; Intelligent Computing (SAIC)</source>
          .
          <source>Igor Sikorsky Kyiv Polytechnic Institute. Kyiv, Ukraine. October 08-12</source>
          ,
          <year>2018</year>
          . P.
          <volume>25</volume>
          -
          <fpage>30</fpage>
          .
        </mixed-citation>
      </ref>
      <ref id="ref27">
        <mixed-citation>
          [27]
          <string-name>
            <surname>Samokhvalov</surname>
            ,
            <given-names>Y.Y.</given-names>
          </string-name>
          <article-title>Development of the Prediction Graph Method Under Incomplete and Inaccurate Expert Estimates / Cybernetics and Systems Analysis</article-title>
          ,
          <volume>54</volume>
          (
          <issue>1</issue>
          ), pp.
          <fpage>75</fpage>
          -
          <lpage>82</lpage>
          . (
          <year>2018</year>
          ).
        </mixed-citation>
      </ref>
      <ref id="ref28">
        <mixed-citation>
          [28]
          <string-name>
            <surname>Litvak</surname>
            ,
            <given-names>B.G.</given-names>
          </string-name>
          <article-title>Expert information: Methods of obtaining and analysis / B</article-title>
          .G. Litvak. - M.: Radio and communication,
          <year>1982</year>
          . - 184 p. (
          <year>1982</year>
          )
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>