I’m a PhD student of Yanka Kupala State University of Grodno.

My specialty is
Mathematical and instrumental methods of economics. The main idea of this specialty is to
develop the theory and methodology of the analysis of economic processes and systems by
using mathematical and instrumental methods. The specialty is aimed to develop mathematical
tools of economic research, methods of improving the validity of management decisions at all
levels of economy and elaboration information technology for economic problems’ solutions
and its expansion into new economic applications.
My PhD thesis is entitled “Modeling and optimization of multi-agent situation analysis’
procedures in decision-making”. The aim of the research is to improve methods and techniques
(models) of multi-agent (multi-criteria) situational analysis for justify management decisions.
Of cource, multi-agent and multi-criteria are not the same. But we were not sure that it
will be possible to make a multiagent system, that’s why we also have a word “multi-criteria”.
I want to start with theoretical results of scientific. In management we clarified the
concept of situational analysis.
The situational analysis is a complex of management tool that includes the identification
and assessment of possible changes in the object’s working. It takes into account the influence
of the present external and internal factors, the possible risks and anticipate developments. The
situational analysis is aimed to the current situation’s transformation on the basis of
development’s prediction and possible risks with using the best methods for goals’
implementation. Then the functions of the situational analysis are data collection, information
sharing, data processing, decision- making and system’s self- evaluation.
Wide functionality needs to appeal several (many) agents to this process. In this case, the
agent may be not only the individual, but a group of individuals or software product that have
the ability of perceive the environment and influence to the object on the basis of this
information. The interaction of agents leads to the multi-agent systems formation. The main
reasons of using the multi-agent theory to the situational analysis’ process are following:
• large number of different criteria, preferences and restrictions;
• the difficulty of prediction appeared as a result of uncertainty;
• the necessity of taking into account occurring events and other managemant decision;
• high efficiency of decision-making;
• limitation of resources (labor, material, computer, etc.);
• experts’ independence and their adequacy within a limited range of issues.
We need to simplify procedures, to form a clear relationship between them and organize
the flexible structures with abilities of learning, cooperation and self-organization. The agent’s
interaction in the system can be represented as follows:
 The advantage of the multi-agent situation analysis’ technique is speed-up of the
procedure. This is due to, firstly, possibility of using agents in the decision-making of different
projects at the same time. Secondly, the method steps may perform in parallel. Thirdly, each of
the agents provides a limited procedures (functions), so they are small and don’t require a lot of
resources for functioning.
The basic principles of the developed concept are following:
- The possibility of viewing situations with varying degrees of detail, their decomposition
into components, a set of elements (layers);
- The one-time data entry;
- The search for solutions to the existing (base) model of the situation;
- Determination of the probability of occurrence of situations (events);
- Systematization of information;
-Flexible classification system (based on the principles of fuzzy clustering);
- The historicity of situations (possibility of tracking the development);
- Visualization of the information;
- Generation of reports;
-self-education of the system.

One area of using of multi-agent situation analysis is its application to the decision-
making for the region governance. Let’s consider the features of region governance. The region
governance can be seen in two aspects. In one side region is seen as an independent system
with its own resources and goal to increase the life’s quality of the local population. In other
side the region is considered as a subsystem of the state (or the higher regional formation). The
aim of region’s functioning in this aspect is to find out resources for national economic growth.
These two aspects relate and condition each other.
The system of local government includes departments and agencies with dual
subordination to the ministry and to the local executive committee. Therefore, the inter-
departmental barriers may lead to a slowdown in decision-making although the linear
submission to the chairman of the executive committee.
We should also take into account the complexity of the issues of regional governance,
their broad functional orientation and territorial distribution.
The situational approach allows us to take a fresh look at the regional administration.
Given the multiagent approach the structure of the system could be as follows:
 We can discribe six main types of agents that must be present in the developed system:
• Agent 1 works with databases of statistical control.
• Agent 2 is responsible of the Bank of situations, determines the degree of non-standard
problem situations, finds out solutions of problem on past experience.
• Agent 3 analyzes the data, determines main factors, works out scenarios. There can be
some several agents. Each of them can use different methods of data analysis and prediction.
• Agent 4 selects the optimum scenario taking into account the limitations by national
programs of development, available resources and potential.
• Agent 5 assesses the coherency of the experts’ (agents’) views.
• Agent 6 is responsible for the security of the entire system.
The using of learning agents allows realizing the procedure autonomously, without the
involvement of a large number of experts.

Why it’s actual and what is differences between our approach and previous :
1) We look to the lowest regional level (rajon).
2) we develop a software for decision-support in regional investment solutions.
3) we collect data from different sources in one database.
4) the system can be used not only by regional and national government for improving the
regional investment attractiveness but also by investors for choosing the most likable region.

My test subject area is management of regional medium-and long-term development on

the example of Grodno region. We applied the theory of situational analysis to the socio -
economic development of Grodno region and systematized the major factors predicting gross
regional product. For this we collected 28 indexes of 17 th areas of Grodno region in 2008-2012.
Based on this data we divide all region to 4 clusters with same level of socio-economic
development. And we proofed that there’s a significant correlation between population and
gross regional product. And the most important factor (except population) is investments.

Conclusion:
In theory
In practice: The results of monitoring the development of Grodno region; the main
factors of socio-economic development of the areas was identified; the method of estimating
the contribution of regions to the result of socio-economic development and its prediction was
developed.
The results are based on usage of mathematical methods, including factor analysis,
cluster analysis, the fuzzy theory and others.
In future we plan:
In mathematical part:
1. to devote characteristics of agents for each step of situation analysis.
The procedure of situation analysis was divided in 6 steps. Each step need a special characteristics of
agents because each agent has own functions.
2. to work out a math model of each step.
We want to make a model of each step. Some of them should have an optimization function and
restrictions. It’s interesting to use a fuzzy theory for reducing experts.
3. to combine all steps in one system.

In information technologies’ part:

4. to produce a software (specification) for each agent.
5. to work out a base of agents’ collaboration.

In practice part:
6. To apply this system to regional governance.
7. To work out a specification of decision support system for regional governance.