AN ONTOLOGY-DRIVEN KNOWLEDGE-BASED APPROACH TO COMPLEX SYSTEMS MANAGEMENT

Abstract

The paper addresses the problem of managing a corporate network as a complex functional system, whose efficiency is determined by the consistency between the infrastructure structure, resource characteristics, and the requirements of application tasks. The feasibility of employing a knowledge-based approach, particularly ontology-based modeling, is substantiated for the formalization of corporate network structure and decision support in its configuration. A conceptual model of a corporate network is proposed, where the network is considered as a multi-level system of interconnected components, including network nodes, services, users, resources, operational parameters, and security constraints. Based on this model, a theoretical foundation for constructing a corporate network ontology is developed, ensuring a consistent representation of infrastructure entities, their properties, and relationships. Furthermore, it enables the formalization of interaction rules and logical constraints in the form of ontological axioms. The study proposes a formalized representation of the corporate network ontology in the form of a tuple-based structure, integrating sets of objects, parameters, states, task performance characteristics, and axiomatic constraints. On this basis, a parameter normalization model is developed, enabling the transformation of heterogeneous characteristics into a unified evaluation scale, taking into account the task context and ontological constraints. This facilitates the construction of an integral configuration quality criterion that incorporates resource, operational, and functional aspects. The main contribution of the work is the development of a method for ontology-driven configuration of a corporate network based on a set of tasks. The method relies on tuple algebra operations to generate a set of feasible configurations that satisfy structural, parametric, and axiomatic constraints, followed by the selection of an optimal configuration using multi-criteria evaluation. The proposed approach ensures transparency, reproducibility, and interpretability of the configuration process. Experimental studies were conducted for two types of corporate networks: a university information and telecommunication system and a network of a commercial enterprise. The results demonstrate that the proposed method significantly improves key performance indicators, including reduced service access latency, decreased computational resource load, fewer access conflicts, and enhanced compliance with security policies. The integral configuration quality criterion decreased by approximately 48–51%, confirming the effectiveness of the proposed approach. The proposed approach can serve as a foundation for the development of intelligent corporate network management systems, as well as for further advancement of optimization methods for complex information and telecommunication systems based on ontology-driven knowledge representation.