STRUCTURED METHOD OF CLOUD INFRASTRUCTURE IMPLEMENTATION

Abstract

Ukrainian enterprises are increasingly adopting cloud services (CS) as the foundation of their businessprocesses. However, selecting the optimal CS configuration has strategic importance, as it ultimately determines the levelof IT infrastructure costs, performance, and scalability of IT systems, as well as data security and regulatory compliance.The decision-making process in this field is extremely complex due to its multi-criteria nature. A decision-maker involved in the selection of CS must usually evaluate conflicting requirements. Classical methods of multi-criteria optimization are not always capable of accounting for subjective priorities and vaguely defined requirements (e.g., “acceptable cost,” “high security,” “sufficient scalability”). Consequently, this reduces the practical applicability of such optimization approaches.The purpose of this study is to synthesize models capable of combining formal quantitative indicators and expert judgments, ensuring a balanced decision-making process.Methodology. The article proposes an integrated model that combines fuzzy logic (FL) with modern evolutionary algorithms for multi-objective optimization (NSGA-III and MOEA/D). To evaluate alternatives, a system of fuzzy membership functions and an aggregation mechanism based on the Mamdani method were applied. This approach enabled the formalization of qualitative and vague criteria. At the optimization stage, a set of Pareto-optimal solutions was generated to represent the trade-offs between different requirements. To enhance interpretability and convenience of selection, the obtained Pareto set was additionally ranked using the Fuzzy Analytic Hierarchy Process (Fuzzy AHP – FAHP) and the desirability function. Overall, the proposed model provides a comprehensive consideration of both the technical-economic parameters of cloud services and the subjective preferences of the decision-maker. The effectiveness of the model was validated through a computational experiment (CE).Scientific novelty. The results of the CE demonstrated improved Pareto front coverage and higher interpretability of decisions compared to traditional optimization methods for CS structure selection in enterprises.Conclusions. The presented model can be used as a decision-support tool in enterprise IT infrastructure management,enhancing the justification and adaptability of cloud service selection.