Subject domain model structure in group modeling

• Date: 2026
• Main Authors: Циганок, В. В., Мінас, В. П.
• Format: Article
• Language: Ukrainian
• Published: Інститут проблем реєстрації інформації НАН України 2026
• Subjects: knowledge transfer platform; hierarchical goal model; collective modeling; group decomposition; expert evaluation; artificial intelligence tools; decision-making support; events development scenarios; strategic planning
• Online Access: https://drsp.ipri.kiev.ua/article/view/363352
• Journal Title: Data Recording, Storage & Processing

This research is devoted to the development and improvement of the structure of domain-specific models created by groups of experts working collaboratively on specialized knowledge transfer platforms. The relevance of the research is determined by the need for analytical support in decision-making within weakly structured domains where the objective function cannot be analytically defined. Unlike classical tree-like hierarchical characteristic of the analytical hierarchy process, using a structure in the form of a network-oriented graph has been proposed. This approach allows for combining sequential "top-down" decomposition with additional relationships identified during the "bottom-up" process, which significantly improves the adequacy and completeness of the description of complex systems. The fundamental basis of the proposed method is the concept of elementary expert decomposition (EED), which is viewed as an autonomous "building block" of expert knowledge. EED consists of a specific goal and a set of sub-goals (influencing factors) formulated by an individual expert, with their number limited by the cognitive threshold of 7±2. To ensure the reliability of expertise, emphasis is placed on the importance of expert anonymity and minimizing unnecessary information in the interface, which helps avoid the "dictator" effect and cognitive distortions. The process of synthesizing a generalized model involves grouping sub-goals that are identical in content using artificial intelligence tools and under the supervision of a knowledge engineer. The study focuses particularly on implementing a mechanism for alternative options to achieve goals by identifying incompatible pairs of sub-goals. This enables the model to adaptively select the most effective decomposition scenario based on current conditions, which is critical for strategic planning and scenario analysis. Practical validation of the developed solutions in the "Consensus-2" web system confirmed the possibility of effectively integrating expert knowledge from various fields, ensuring high-quality recommendations from decision support systems. Fig.: 3. Refs: 8 titles.