The article presents the structure of intelligent decision support systems based on situational center. The structural and functional diagram of the situational center is given, the module for accounting for uncertainty is supplemented. The intellectual subsystem of the situational center is described and the conclusion is done on the applicability of integrated software that allows solving various problems based on different sources of information, taking into account uncertainty. The block diagram of the algorithm for accounting for the uncertainty of the initial information is presented, for which it is necessary to calculate the following sample parameters: size, noise, and outliers, stationarity, type of distribution law. To implement such an algorithm, a software module was designed in the Python programming language, the results of the module are used in all subsystems of the intelligent situational center.
The solution of applied problems requires the improvement of existing approaches to information and analytical support of management activities and the necessity for methodical software and hardware tools. An effective form of information and analytical systems that combining these tools are situational centers (SС), which are increasingly used as a tool to support management activities.
Situational centers, unlike traditional systems of automation of management, enable
in real-time not only to provide the most complete and operatively information on the
current and retrospective situation but also to calculate and analyze the consequences
*
of management decisions [
To solve these problems, it is necessary to include in the SC a separate of relevant
subsystems for monitoring, modeling, planning and forecasting, and decision-making.
[
A great place in the work of the situational center is occupied by solving tasks related
to monitoring, analysis, and assessment of the situation:
monitoring information of various formats, heterogeneous and fragmented in
composition;
operative tracking of a single information space, in which all the main services of
the situational center work;
control and operational impact based on timely information;
identification of a set of factors that have the greatest impact on the development of
the problem. [
The following items relate to planning tasks within a situational center:
the formulation and definition of a system of indicators that influencing the
problematic situation or the decision process as a whole;
justification of the indicators of the proposed strategies, goals, and tasks, which are
planned in the coming period;
definition of the requirement, planning of volumes, and structure of necessary
resources[
The process of forecasting the development of facilities in various situations is an
integral part of the work of any large system that is part of the situation centers, both state
and commercial organizations. So, there arises the problem of choosing an acceptable
control for large objects with a large number of significant parameters and connections
between them. The main tasks to be solved in the forecasting process are the following:
predicting indicators of different purposes and formats;
formation of integrated integral estimates;
fulfillment of target forecast calculations taking into account various parameters;
justification of significance and assessment of the attainability of development goals
[
In the decision-making process, the main goal of the functioning of the situational
center is to solve the set problem under conditions of uncertainty and heterogeneity of the
information, and the following main tasks are considered:
complex solution of the problem based on formal and informal decision support
methods;
generation of maximum possible solutions;
selection, quantitative and qualitative evaluation of performance criteria;
selection and optimization of the solution. [
These subsystems realize the solution of the tasks standing before the situational center. However, in addition to them, it is necessary to include information subsystems supporting the integration of software and hardware into the SO structure.
The situation center provides an opportunity to quickly respond to emerging problems and situations. At the same time, depending on the functions and tasks facing the SC, three operating modes can be distinguished: normal (development of scenarios for the development and functioning of the system), planning (allows you to find hidden cause-effect relationships of processes, predict), crisis (necessary for making decisions in complex or emergencies).
The analysis of literature sources allows us to correlate the operating modes of the
situational center and its functions (table 1) [
From Table 1 it can be seen that the forecasting problem is solved in each mode of operation of the situational center. The decision-maker uses the information obtained as a result to form effective decisions. A large number of different algorithms are used for forecasting, which must be selected depending on the operating mode of the SC, the degree of uncertainty of the source information, the requirements for the forecast and input data.
Forecasting is one of the basic functions of the SC. The decision-maker relies on the forecast to form effective solutions, which allows us to speak about the importance of this task. To implement it, a large number of different algorithms are used, including intellectual ones. The choice of a specific algorithm for solving this problem is required to be performed depending on the operating mode of the SC, the degree of uncertainty of the source information, the forecast requirements, and the input data. The prediction algorithms can be implemented as separate program modules. However, for the successful solution of the tasks in the SC it makes sense to implement a single module, in which it is possible to connect submodules that implement one or another algorithm. These submodules must be connected and used as needed and depending on the requirements of the decision-maker, the input data, the mode of operation of the situational center. Also, the existence of a single forecasting unit makes it possible to establish the process of transferring data to other modules, this is achieved through a single data format and processing of information.
A preliminary analysis of the available information is necessary during the operation of the SC to determine the type and degree of uncertainty present in it, to apply methods for its accounting and minimization. The need to account for uncertainty is justified by the following facts: study of poorly structured systems that use both quantitative and qualitative characteristics, with the latter prevailing; system management is carried out with incomplete, insufficient knowledge about the state and influence of the environment on the system, especially with large investments of resources; the presence of isolated or conflicting subsystems, while the search for a solution can be reduced to finding some compromise; when managing complex systems, it is more rational to generate proactive solutions than corrective ones.
Work with information is carried out at all stages of the SC. The operation of the subsystems of the SC is based on the work with data of different origin and structure. In this connection, the issue of storing heterogeneous sources of information used for analysis and decision-making is actual. To do this, we use a universal data warehouse, which allows us to combine a large number of information resources using both standard databases, and with the help of new cloud technologies.
To implement such an analysis, a methodology for accounting for the uncertainty of
the initial information was proposed in [
From the point of view of the system approach, the decision-making process does
not depend on the area in which the decision is made. At the same time, it is necessary
to provide maximum support for monitoring, forecasting, planning, and
decision-making processes from the point of view of mathematical methods capable of providing
intellectualization of the stages of functioning of the situational center. For this purpose,
a methodology for organizing and using mathematical methods and algorithms for rapid
and effective analysis and solution search has been developed for the situational center
[
The module is based on an algorithm that allows you to evaluate and determine the set of parameters of the source information, prepare data for further processing in an expert system that selects the most suitable type of uncertainty. 3
The obtained type of uncertainty allows us to assume the preferred form of the mathematical model for describing the system; for this, it can be used both previously stored information and additional information about the problem received from the decisionmaker.
The structural diagram of a software module for identifying and accounting for the uncertainty of the source information, taking into account the technologies used, is presented in Fig. 3.
Within this scheme, the following structural blocks can be noted:
Sampling. In this block, a selection is obtained with information about the situation,
object, or phenomenon. Data can come from various sources - from the user, from
the Internet, etc. For their initial processing, external monitoring tools and intelligent
search can be used.
Interface. This unit can interact with the user through a graphical user interface
(GUI), and an external application through standard data types. It should also be
noted that it is necessary to provide for the functioning of the system in two modes:
─ Via the Internet. For this, in the Python programming language, there is a Django
web framework. This framework is free and uses the
Model-Presentation-Controller (MVC) design pattern, and allows to build a system from separate, plug-in
modules.
─ In the form of an application program (native application), which is designed for
a specific operating system. To create a GUI in this case, it is advisable to use
PyQt. This solution has a large number of ready-made elements for creating a
graphical interface, supports rapid prototyping tools.
Data preparation. In this block, the input sample is read and converted to the internal
data format. Here, the basic processes for sorting data, their initial processing are
carried out. Numeric tables are also manipulated. For this, the Pandas library was
chosen - a library in the Python programming language for processing input data. It
provides the ability to interact with multidimensional structured data in CSV format.
There is the possibility of obtaining data slices (work not with the entire sample, but
with its part).
In the unit for calculating the characteristics, the main parameters characterizing the
uncertainty are determined: size, availability of qualitative data and dates,
determination of the law of distribution of a random variable, identification of noise and
emissions, calculation of the scale and scales of numerical data, as well as the
determination of the stationarity of the process. The following libraries of the Python
programming language were used to carry out these mathematical calculations:
NumPy (for working with matrices and arrays) and SciPy (for basic mathematical
calculations)[
The calculated type of uncertainty of the initial information and the mathematical model is then used by other subsystems of the situational center. This approach can operate in an automated mode, which significantly speeds up the decision-making process and reduces the requirements for user qualifications.
As a platform for building an intellectual system of a situational center, it is advisable to use integrated software that can solve various tasks based on heterogeneous sources of information with full uncertainty.
The presented basic functional characteristics of systems, objects, and subjects of the decision-making process allow considering the complex approach to the realization of the intellectual system of the situational center. The use of such an approach makes it possible to provide the possibility of considering and solving the maximum number of tasks, reducing the time for analysis and preparing information for the solution, using an intellectual approach to the extraction and use of diverse knowledge. Also, it should be noted that the functioning of all structural components should be accompanied by mandatory consideration of the uncertainty of available information. 4
The use of different systems, including software and hardware systems, decision-making subjects, mathematical and heuristic decision-making methods within the single intellectual system of the situational center, will ensure the fulfillment of a large number of tasks involving a minimum amount of resources.
The proposed structural and functional scheme of the intellectual system of the situational center ensures the implementation of the most effective strategy for finding the optimal solution to the problem or task in question.
The use of such an approach will provide intellectual support for monitoring, planning, forecasting, and decision-making procedures for functioning in different modes at any stage of the SC operation. The developed intellectual decision support system based on the situational center is a set of interrelated software systems, mathematical and methodological apparatus, decision-making specialists that allow the intelligent system to be reconfigured to solve a different class of tasks and different subject areas in the shortest possible time.