Method for optimizing cybersecurity strategies in distributed information systems based on NSGA-II multi-objective optimization and game theory

Abstract

This paper presents a method for optimizing cybersecurity strategies in distributed information systems based on NSGA-II multi-objective optimization and game theory. The relevance of the study is driven by the growing number of cyber threats, the increasing complexity of distributed information system architectures, the continuous growth of processed data volumes, and the need to improve decision-making efficiency when developing cybersecurity strategies under resource constraints and dynamically changing cyber environment conditions. The study analyzes current approaches to ensuring cybersecurity in distributed information systems, risk assessment methods, multi-objective optimization techniques, and the application of game theory to strategic cybersecurity planning. It is established that existing approaches do not sufficiently consider the simultaneous influence of risk level, implementation cost of security mechanisms, effectiveness of cybersecurity measures, and potential attacker behavior when forming optimal cybersecurity strategies. A mathematical model for cybersecurity strategy optimization is proposed, combining multi-objective evaluation of alternatives using the NSGA-II algorithm with a game-theoretic model of strategic interaction between the cybersecurity system and the attacker. A system of optimization criteria is developed that takes into account risk level, costs of implementing security measures, the level of information resource protection, and incident response time. The architecture of a decision support system, the operational algorithm of the proposed method, and a procedure for generating a set of Pareto-optimal strategies are presented. A computational experiment was conducted for different operating scenarios of a distributed information system. The simulation results demonstrated a reduction in the overall risk level, improved effectiveness of cybersecurity strategies, enhanced system adaptability to changing operating conditions, and advantages over traditional approaches. The obtained results confirm the feasibility of combining NSGA-II multi-objective optimization and game theory to support decision-making in the field of cybersecurity.