ADAPTIVE BIG-DATA MANAGEMENT OF SMART RETAIL ENTERPRISES

Abstract

The paper develops and substantiates an information technology for the adaptive management of a smart retail enterprise based on Big Data. The proposed solution is aimed at creating an integrated closed-loop management environment in which heterogeneous internal and external data streams are transformed into knowledge, forecasts, managerial decisions, and corrective actions in near real time. The study formulates a formalized system of functional, architectural, and operational requirements for such technology. These requirements include adaptability, data integration, scalability, support for batch and streaming processing, low decision latency, data security, fault tolerance, service orientation, and the capability of continuous self-learning. The methodological foundation combines the systems approach, the cybernetic approach, and data-driven management principles with methods of data mining, machine learning, forecasting, optimization, and multicriteria decision making. A formal structure of the technology is proposed as a set of interconnected subsystems for data collection, integration, storage, analytics, forecasting, decision making, implementation, and monitoring. Their interaction is described as a closed transformation cycle that links data acquisition with feedback-based managerial correction. The paper further develops the structural representation of the technology by distinguishing three completed functional stages: data formation, analytical processing with knowledge generation, and managerial decision implementation. A particular contribution of the study is the formal consideration of the relationship between the data-flow intensity and the throughput of the computing infrastructure. This makes it possible to define a real-time operation condition and to explain how overloads, queues, and excessive delays can be prevented when processing high-volume and high-velocity data streams. To assess the performance of the proposed information technology, a system of partial and integral criteria is introduced. The integral multiplicative efficiency criterion jointly takes into account qualitative, temporal, and resource parameters of all stages of the management cycle. In addition, an adaptability criterion is proposed to evaluate the quality of system response, reaction speed, and resource expenditure under changing operating conditions. The obtained results provide a formal basis for designing scalable intelligent management platforms for smart retail enterprises, improving the consistency of information processes, increasing the quality of forecasts and managerial decisions, and supporting proactive enterprise behavior in a dynamic digital economy.

Keywords: adaptive management, smart retail enterprise, Big Data analytics, intelligent decision support systems, machine learning, real-time data processing, data-driven management.