| Description: |
The article considers the current scientific problem of multi-criteria optimization of the information protection systemat critical infrastructure facilities, where conflicting requirements are simultaneously present: the need to minimize risks, costs,and delays while ensuring high availability and speed of response to cyber threats. The application of the weighted sum ofcriteria method is proposed as an effective tool for reducing the vector objective function to a scalar form, which allowsintegrating the subjective priorities of the infrastructure owner through setting weight coefficients. Special attention is paid tothe conditions for the correct application of the method, in particular, the assumption of the convexity of the Pareto front, whichguarantees obtaining a complete set of Pareto-optimal solutions. The need for pre-processing of criteria is justified - unificationof the optimization direction, normalization of values measured in different units, and formalization of constraints related to thebudget, latency, and availability level. Using the example of three key criteria – risk, cost and response time – the constructionof an objective function is demonstrated, which is minimized within realistic constraints typical for large infrastructure facilities.The analysis of existing approaches (ε-constraints, AHP, TOPSIS, evolutionary algorithms) confirms the advantages of theproposed method in the context of rapid prototyping and practical implementation due to its computational simplicity,interpretability and compatibility with classical optimization methods. The results of the study emphasize the importance of theweighted sum method as a scientifically sound and practically applicable approach to designing effective, economically feasibleand functionally stable cyber protection systems for critical sectors of the economy.Keywords: information protection system, critical infrastructure facilities, multi-criteria optimization, weighted summethod, Pareto optimality, cybersecurity, risks. References1. ... |