Abstract
A software implementation of the problem of multi-criteria optimization of the selection of information protection means (IPS) for a university computer network has been developed based on the evolutionary algorithm NSGA-II. A cybernetic model for selecting the optimal set of IPS has been considered taking into account a set of criteria, including the cost of implementation, the level of reliability, and coverage of current threats. Optimization objective functions have been determined, each of which is a function of discrete variables described in the literature and characterizes possible IPS configurations. In the software implementation of NSGA-II, an algorithm for generating the initial population of solutions with crossover, mutation, and selection mechanisms was implemented, which provides an effective search for Pareto-optimal configurations, and computational experiments were also conducted to demonstrate the influence of model parameters on the optimization result visualized on graphs. A Pareto front was obtained, which visualizes the trade-offs between cost, reliability, and level of protection. The results obtained during the study generally confirm that the use of the evolutionary algorithm NSGA-II allows achieving balanced solutions when designing an information security system for a university network. The software implementation of multi-criteria optimization of the ISI parameters for a university network, presented in the article, is implemented in the Python language using specialized optimization and data analysis libraries, ensuring the reproducibility of computational experiments and the ability to adapt the algorithm for various scenarios of protecting distributed computing systems.