This article examines an optimal control problem focused on optimizing the structure of a social network to achieve a desired opinion distribution in the population within a finite time horizon. The opinion dynamics adhere to the SCARDO model, and the network structure is operationalized using a stochastic block model whose parameters are subject to adjustment. We derive an analytical result demonstrating that the problem under consideration can be reduced to a control problem in which the structure of the network is fixed, but the parameters of the ranking algorithm—integrated into the model—are optimized. The latter problem is already solved in the scholarly literature. This allows us to introduce a numerical algorithm for solving the initial control problem. Our findings have potential applications in designing friend recommendation systems for real-world social platforms.

This paper considers the problem of automatic answering complex scientometric questions formulated in natural language (NL) over knowledge bases. The study is topical due to the limitations of modern large language models (LLMs): despite high understanding capabilities, they tend to generate inaccurate responses to user questions and may have outdated information in specialized subject areas. At the same time, knowledge graphs provide accurate and relevant information but require knowledge of a formal query language. A hybrid architecture-based solution is proposed: an LLM acts as an intelligent interface to an ontology-driven knowledge base, converting NL questions into correct SPARQL queries, and the results are returned to the user. To solve the problem, a specialized data corpus is compiled to train and test NL-to-SPARQL models in the field of control theory. The approach is implemented based on the ontology of scientific activity in the field of control theory and validated on the generated corpus of questions. Integration of the LLM with the ontology-driven knowledge base ensures a high accuracy of answers (about 99%), which confirms the prospects of this approach.

This special issue presents extended texts of selected reports from the XII International Conference “Integrated Models and Soft Computing in Artificial Intelligence” (IMSC-2024), which was held in Kolomna, Russia, on May 14–17, 2024. The conference was co-organized by the Russian Association of Artificial Intelligence, the Association of Fuzzy Systems and Soft Computing, the Federal Research Center “Computer Science and Control” of the Russian Academy of Sciences, the Administration of the Kolomna Urban District, the Kolomna Institute of Moscow Polytechnic University, and the Institute of Computer Technology and Information Security of Southern Federal University. The scientific topics of the IMSC- 2024 conference covered the following relevant areas in artificial intelligence: fuzzy models, soft computing, measurements and evaluation in artificial intelligence; machine learning and neural network technologies; bioinspired approaches; cognitive models in artificial intelligence; data mining, knowledge and ontology engineering; hybrid intelligent systems; and intelligent agents, cyber–physical systems, intelligent manufacturing.

This paper examines games on networks with linear best responses, which allow for the analysis of how interaction structures influence agents’ strategic behavior. Special attention is given to intervention issues in such models, particularly in selecting optimal intervention strategies aimed at maximizing the central planner’s objective function. Two main control policies are analyzed: individual agent incentives and modifications of the interaction structure. The concept of a representative agent is introduced to simplify equilibrium analysis and control problems in games on networks. Both aggregate outcome maximization problems and adversarial scenarios between competing central planners are considered. Analytical conditions are derived to determine whether controlling the interaction structure is more effective than influencing individual incentives. Numerical experiments confirm the theoretical results and demonstrate their applicability to different types of network structures.

This paper is devoted to the fragility analysis of PI controllers. Two different approaches are proposed to estimate the fragility of a given PI controller; they can be used independently of each other. The first approach is based on the ideas of ellipsoidal estimation and the technique of linear matrix inequalities. Within the second approach, involving the so-called parameter loop opening, the parameter under study is “extracted” from the plant–controller system, forming a fictitious control loop; after opening, this loop is analyzed by classical frequency-domain methods (the Nyquist criterion and D-partition). The features of both fragility analysis methods are described, and numerical examples are provided.