The thesis research is devoted to solving the scientific and practical task of further developing and integrating methods of modeling domain-specific branches of information technology based on cognitive maps and domain-specific programming languages. Scenario analysis processes are chosen as the research subject.
Managing everyday human activity in conditions of growing competition, increasing complexity of tasks, and high dynamics of supply and demand changes require solving work organization problems, monitoring, resource control, and planning. The methodology of scenario analysis based on the domain-specific branch model is used more and more often to solve such problems. Analysis of existing approaches and practical results in scenario analysis shows that cognitive and domain-specific programming integrated with combined information technology (IT) is not fully used to solve this problem. Although individual works propose the use of cognitive maps, there is no information about the means of obtaining multiple scenarios on their basis. As for the problem-oriented languages (POLs), in the scenario analysis, there are some works where POLs are proposed only to describe the scenarios themselves without implementing the possibilities of their generation and processing. Therefore, creating the corresponding POL for generating a set of scenarios and their processing as a part of IT for scenario analysis is an urgent task.
The thesis research aims to improve scenario analysis tools by applying the cognitive model of the domain-specific branch to generate multiple scenarios by creating and applying domain-specific programming languages.
For the first time, IT for scenario analysis is proposed, which, unlike other approaches, allows the generation of multiple scenarios based on the cognitive model of the domain-specific branch.
For the first time, a depth-first search algorithm on a graph has been applied to generate multiple scenarios based on the cognitive map, which, unlike the cross-impact method, generates scenarios only with interrelated events and does not allow the generation of scenarios with mutually exclusive events;
Mathematical modeling tools based on graph structures were further developed by integrating the language structures of mathematical modeling packages and graph databases, which made it possible to expand the functionality of operations with the graph database and, thus, provided an opportunity to create models of the scenario analysis process;
Methods for multi-criteria optimization have been improved due to developing the POL for determining the target function, which provides a simplified process for finding the optimal scenario;
Methods of transforming the relational data model into a graphical one have been further developed by implementing the function of automated POL query generation, which automates this process.
The solution to the problem of creating ІТ of scenario analysis based on the domain-specific branch model requires two separate subsystems: expert evaluation and modeling with scenario analysis support. Modern tendencies of IT development and modeling features make it necessary to create a specialized programming language that is based on mathematical methods and allows specialists without profound knowledge of programming languages to describe domain-specific branches.
In the form of models and information technologies for solving the problem of modeling domain-specific branches for scenario analysis, the thesis research results can improve scenario analysis tools by applying the cognitive model of the domain-specific branch to generate multiple scenarios by creating and using domain-specific programming languages.
The developed models and information technologies were used in NEON Telecommunication Co. LLC (Cooperation Agreement of 01.09.2020). Also, the result of the study is implemented in the educational process of Petro Mohyla Black Sea National University in teaching the discipline "The theory of automata and formal languages" for third-year students of the specialty 122 - Computer Science.
