Fuzzy identification of the state of IT corporation

The methods and approaches to solving the problem of corporate governance are analyzed. Economic, social-psychological, administrative methods of corporate management are considered. Situational, systemic, situational fuzzy approaches to corporate management are considered. It was decided to apply a situational fuzzy approach to solving the research problem. It is proposed to use the k -means clustering algorithm in order to distribute fuzzy situations by classes. Clustering problem is solved. The task of identifying the state of IT corporation. An example has been calculated with which you can conduct a comprehensive study of the state of the object, which allows you to get a full understanding of the estimated IT corporation not only at the time of the assessment, but also in the retrospective and forecast periods.


Introduction.
Currently, the evolution of the community is largely dependent on IT technology.The IT industry is one of the four priority sectors for Ukraine's export strategy.In recent years, there has been an increase in the number of IT corporations in Ukraine, which are rapidly expanding.The creation of an effective ITtechnology requires the creation of an optimal ITcorporation management system.Thus, the task of managing the IT corporation becomes urgent.
The modern concept of "management" means the process of coordination of various activities taking into account their goals, conditions of implementation, stages of implementation.
Considering the management process as an element of the organization, it is necessary to determine what tasks the management of a specific organization performs as it does, that is, to understand which sectors are the subject of coordination and regulation and in what way this coordination is carried out.
The Corporation has the following characteristic features: nonstationarity of the process occurring in the system, a large amount of information, the impossibility of constructing a mathematical model from the standpoint of the classical control theory.The solution of the management task stipulates the following tasks: aggregation of information, identification of the state of the system and solution of the management task.
The present work is devoted to aggregating information and solving the task of identifying the state of IT-corporations.It is proposed to use the elements of situational control and the mathematical apparatus of the theory of fuzzy sets.
Since a large amount of information complicates the construction of the situational management model, it is proposed to aggregate it using cluster analysis.
Analysis of the research problem.The quality of IT-corporation management is largely determined by the effectiveness of decisions taken by leaders of different levels -the higher, middle and lower levels.To make such decisions, managers need to have reliable, timely and regular information about the state of affairs [1].
Particularly acute problems of reliability, timeliness and regularity of management information face large IT corporations.Information systems are the only way to collect, process and present data in a form that is convenient for making managerial decisions.
It can be concluded that in modern conditions ITcorporations need a tool that allows using the available resources as efficiently and economically as possible.The use of foreign experience, concepts and technologies can be of great use only if it is really a progressive experience and if it is used taking into account the specific conditions in which a particular IT corporation is located.Management systems (methodologies, approaches, computer programs) should be flexible enough so that they can be adapted as needed, with various changes in the "external environment".
Thus, it is timely to introduce into the IT-corporation a management system that could make its work more efficient and, as a result, lead to an increase in profits.
Description of the research object.In the work as an object of research is considered an IT corporation, which represents the domestic software industry in the world market.
At the head of the IT corporation is the General Director.He represents it in all institutions and state organizations, concludes contracts, issues orders for the corporation, opens accounts in banks and performs many other functions.
IT-corporation has a linear-functional (combined) management structure, because it is based on a close combination of linear and functional connections in the management apparatus.It provides such a division of labor, in which the linear links make decisions and manage, and the functionalconsult, inform, coordinate and plan economic activities.The basis of the organization of functional actions is based on a linear principle.This is illustrated by Figure 1.The linear-functional structure is also characterized by weak horizontal connections between functional divisions.Therefore, quite often some similar control functions are not coordinated enough.The constant need to harmonize decisions at the level due to the variety of horizontal links causes a significant slowdown in the timing of the implementation of the goals, a decrease in the quality of decisions made, an increase in management costs.
It should be noted that the main criteria for the effectiveness of the IT corporation can be its indicators: the profit that the corporation receives for one quarter (hryvnia), the number of timely completed orders (pcs per quarter), costs, per order (man-hours).
All indicators of IT-corporation activity tend to grow, and the growth rate of profit from sales of products and services exceeds the growth rate of sales proceeds, which in turn exceeds the growth rates of prime cost by a favorable trend.
To solve the management problem, it is proposed to use an unclear situational approach.Since the IT corporation has an established, mature production: a largescale, mass.The goals here are clear and rarely change.The main thing is to be able to exit in the best way from the unfavorable situations that are created, to avoid failures in the production process.
The management decision always consists in the choice of certain quantities, which are called "management".They, in turn, determine the values of other quantities -the phase coordinates.
To solve the research problem, the following phase coordinates: 1 x -number of departments in the IT company (pcs); Due to the large amount of information, it is necessary to aggregate it using fuzzy logic [2].Each attribute is formalized by a linguistic variable.Next, we set the base set for the characteristic.So, every sign , where To formalize fuzzy situations of the IT-corporation management process, the vector of phase coordinates: 3 T = {"few", "enough", "much"}; 5 T = {"few", "enough", "much"}; 5 D = {30 000; 100 000}.Next, we form the control vector.For the considered domain, it has the form: 1 u -number of employees in IT-corporations (IT- specialists); <"Number of employees in IT-corporations",  1 T = {"few", "enough", "much"}; 4 T = {"few", "enough", "much"}; 4 D = {10; 100}.The criteria for assessing the functioning of the corporation are as follows: 1 K -profit that the corporation receives for one quarter (in UAH); <"Profit that the corporation receives for one quarter", 1 = {"few", "enough", "much"}; 1  3 T = {"few", "enough", "much"}; 3 D = {1000; 8000}.The process of functioning of the system reflects the behavior of the system in time and can be represented as a successive change of its states.If the system changes one of its states to another state, then it is customary to say that the system goes from one state to another.
Statement of the research problem and the choice of the method of its solution.The paper considers the ITcorporation, which has a linear-functional management structure.In modern development this organization is one of the leaders in software development.IT-corporation has established, fulfilled and large-scale production.Effective activity of the corporation is satisfactory, however management believes that the IT corporation has a great potential for organization.
Thus, to solve the an IT corporation, first of all, it is necessary to solve the problem of identifying the state of the corporation and to develop measures to ensure its further effective development.
So, the IT-corporation's need to redefine target targets and their implementation can be met by creating flexible tool-and-methodical support, including two main processes: information aggregation, IT-corporation status identification.
Thus, the first stage is reduced to the transformation of the detailed information into "packages" (aggregates) of data that allows to analyze the IT corporation in terms of a small number of relevant aggregate variables that include profit, costs, orders (intermediate and final), the number of employees in the corporation, the amount of investment, the amount of technical support, etc., under conditions for which the initial data are known with varying degrees of detail.
The second stage is the identification of the state of the IT corporation.It consists in determining from the input and output values of such a situation from a particular class of situations, in which the real system under investigation is equivalent..In accordance with this, it is necessary to define a class of situations models, among which the most suitable (current) model of the situation will be chosen.That is, it is necessary to conduct a comprehensive study of the state of the object, which allows to get a full understanding of the estimated IT corporation not only at the time of the assessment, but also in the retrospective and forecast periods [2].
The management of the corporation should be able to implement its proposed management in any period (once a week, once a month, once a quarter).Then compare the results that were before the introduction and the results were obtained after the innovations.Thus, the IT corporation for 2 years should reach an entirely new level of management, which would lead the corporation to achieve its main goal.
Economic, social-psychological, administrative methods of corporate management are considered.Situational, systemic, situational fuzzy approaches to the management of the IT corporation.
It is revealed that the application of the above methods, as well as the situational approach, is ineffective for our research task.So there are many signs that are quite difficult to describe the situation in the IT corporation.All this is a complex mathematical model.The authors decided to switch to fuzziness.Namely, it was decided to apply a situational fuzzy approach to solving the research problem.
The task of clustering.There are data on the situation describing the state of the corporation.In cluster analysis this is called observation.Each situation is described by signs, for example, the number of technologies for creating software products, the number of funds allocated for a particular project, the number of reserve personnel, the number of employees who own certain technology and the number of departments in the corporation.They form a vector representing the situation.This vector has the form of a list of numbers that can be interpreted as coordinates of a multidimensional space [3].
In order to distinguish classes of situations using the fuzzy k-mean method, we use the MatLab application package for numerical analysis.
In the Statistics Toolbox section, there is a standard function kmeans(), access to which allows to obtain clustering results for the sample.The input data for it are sampling (value k), as well as numerous parameters and settings.
Authors were given the following data for the calculation of the test case: k = 3, ka set of characteristics that adequately reflect the properties of the research object (Table 1 The data given in table 1 form 3 classes ("few", "enough", "much") to describe each of the pranks characterizing the corporation.These vectors are represented in the form of a list of numbers in Table 2. Next they will be interpreted as coordinates of the multidimensional space for the MatLab.
The set of situations formed by the authors is partially given in Table 2.The initial data are heterogeneous, and their estimates by parameters are not consistent, which corresponds to the real situation.
We determine the value of the proximity measure for each criterion.Let us write the obtained values for characteristic 1 x in Table 3.

Distance between signs
The value obtained As you can see, the smallest value between the second (the number of technologies for creating software products) and the fifth (the amount of funds allocated to a particular project).And the greatest between the third and fourth signs, as well as between the fourth and fifth.
Next, a pair of numbers is formed, which consists of the value of the characteristic and the value of the membership function.Then we get the formed situations.
Let's test the test case.Let the current situation б will be formulated as follows: IT-corporation has 4 technologies used to write IT solutions and 3 employees with a certain technology, that is, the total number of developers is 12 people.Also now, the IT corporation is developing a low-budget project (40 000 hryvnia).
The initial data are heterogeneous, and their estimates by parameters are not consistent, which corresponds to the real situation.
We determine the value of the proximity measure for each criterion and write the results in Table 3.
Next, a pair of numbers is formed, which consists of the value of the characteristic and the value of the membership function.Then we get the formed situations.
Let's test the test case.Let the current situation will be formulated as follows: IT-corporation has 4 technologies used to write IT solutions and 4 employees with a certain technology, that is, the total number of developers is 12 people.Also now, the IT corporation is developing a low-budget project (40 000 thousand hryvnia).
Let's write the current situation сas follows: there is a small number of technologies for creating software products, a large number of employees who possess certain technology and a small amount of funds allocated for a particular project.Situation characterized by signs 2x -number of technologies for creating software products, i x 3the number of employees who possess certain technology, l x 4amount of funds allocated for a particular project.Thus, we determine that the characteristic 2 x belongs to the class of situations {few}, We have the situation 0 S is formed as follows: 0 S = {<0,25/"few">, <0,9/"enough">, <0,35/"few">}.Further it is necessary to identify the current situation.
The main advantage of the algorithm is its simplicity.Simplicity usually means high execution speed and efficiency compared to other algorithms, especially when working with large data sets [3].
The two main disadvantages of the method in the medium are invested in the sensitivity to explosions and the initial choice of centers of gravity.
The task of identification.To identify the state of the IT corporation, it is necessary to consider the situation in which it is at the moment.It is necessary to compare the input fuzzy situation 0 S with every fuzzy situation } , , { For a test render, let's use the example described above.The authors identified 22 reference situations.All reference situations are assigned to three levels of the hierarchy.Using the formula, we will determine the degree of inclusion of the current situation in each of the available.To do this, let us take the part formulated by the authors, earlier situations [4].
To the type of fuzzy situations S  include situations that characterize the state of performance of the functional responsibilities of the corporation.
Situations S  are determined by the following set of characteristics (factors): For the test miscalculation, a stack of situations is necessary that belong to the situation S  , because it is characterized by signs  1276) 87 calculation of the degree of fuzzy inclusion and calculation of the degree of fuzzy equality of situations, it can be concluded that at the moment the corporation has a large increase in the number of personnel and sufficient profit for one quarter [5].Thus, it is possible to identify the situation in the IT corporation, which can occur at any time and will be considered current.With the help of this identification, the management of the IT corporation at any time can receive reports on the state of affairs in the corporation, as well as obtain the necessary management decision to achieve its specific objectives.
Conclusions.It has been established that the most effective approach to the management of an IT corporation that has an established, established large-scale production is a situational fuzzy approach.
Since it is necessary to process a large amount of information, it is suggested to use the k-means clustering algorithm in order to distribute fuzzy situations.
Data aggregation is performed.Namely, information was converted into data aggregates, which allowed the IT corporation to be analyzed in terms of a small number of relevant aggregate variables.
The state of the IT corporation was identified.Namely, it is determined by the input and output values of such situations from a certain class of situations, in which the real system under investigation is equivalent.An example has been calculated with which you can conduct a comprehensive study of the state of the object, which allows you to get a full understanding of the estimated IT corporation not only at the time of the assessment, but also in the retrospective and forecast periods.
Then, using the results obtained, we propose the solution of the control problem.

Figure 1 -
Figure 1 -Linear-functional control structure set of a linguistic variable (or a set of linguistic values of the trait; mnumber of characteristic values); i D -basic feature set i x .

2 x
-number of technologies for creating software products.<"Number of technologies for creating software products.

4 u
number of orders of IT solutions (pcs per quarter); <"Number of orders of IT solutions",
the class of situations {few}.

S
measure to determine the degree of proximity of the fuzzy situation of a fuzzy situation 0 S the degree of fuzzy inclusion of a fuzzy situation will be used 0 S in an unclear situation i S .Enabling the fuzzy ISSN 2079-0023 (print) Системний аналіз, управління та інформаційні технології ISSN 2410-2857 (online) 86 Вісник Національного технічного університету «ХПІ».2017.№ 55 (1276) situation 0 S in an unclear situationi The definition of the degree of inclusion of situations, based on determining the degree of inclusion of fuzzy sets.Degree of inclusion of

2 x 1 S 1 S
-number of technologies for creating software products (pcs); includes fuzzy situations that characterize the quality of management in a corporation, and also depend on the profit of an IT corporation for one quarter.Fuzzy situations / few 0,45 / enough 0,9 / much 2 x 0,25 / few 0,45 / enough 0,9 / much i x 3 So, the situation 11 S хcharacterizes the small profit of the IT corporation for the quarter.Situation 12 S characterizes the sufficient profit of the IT corporation for the quarter.Situation 13 S characterizes the big profit of the IT corporation for the quarter.Situation 21 S reflects an insignificant degree of training in new technologies.Situation 22 S reflects a significant degree of learning new technologies.

SS
of situations S  , because situations S  are determined by the following set of characteristics (increase in the number of personnel in the corporation and a sufficient profit for the quarter.Let us now calculate the degree of fuzzy equality of situations.The degree of equality of two fuzzy subsets sets are mutually indifferent.Based on the foregoing, it is necessary to determine the degree of fuzzy equality of the situation 0 and the situation 12 S fuzzy equally.So, based on the received calculations, namely, the ISSN 2079-0023 (print)ISSN 2410ISSN  -2857 (online)   (online)   Системний аналіз, управління та інформаційні технології Вісник Національного технічного університету «ХПІ».2017.№ 55 (

Table 2 -
Developed situations